⚛️ Quantum Computing in 2025: $1.25B Invested This Quarter, But Can It Actually Do Anything?

product_david · October 9, 2025, 9:48pm

Just attended the quantum computing sessions at SF Tech Week and I’m both excited and deeply skeptical. The hype vs. reality gap is MASSIVE.

Sessions attended:

Google “Quantum Computing: From Lab to Reality” (presenting Willow chip)
IBM “The Road to Fault-Tolerant Quantum” keynote
a16z “Investing in Quantum: Separating Signal from Noise”
Panel: “When Will Quantum Actually Matter for Your Business?”

My background: Product lead at fintech startup, trying to understand if quantum is real or just the next blockchain-level hype cycle

The Investment Boom (It’s Real)

Data from a16z session - quantum funding 2025:

Q1 2025 alone:

$1.25B raised by quantum startups
That’s 128% MORE than Q1 2024 ($550M)
70% of 2024’s full-year total… in just 5 months

2024 full year:

$1.9B across 62 funding rounds (138% jump from 2023)
Government commitments: $3.1B
Venture capital: $2.6B

Major rounds:

Quantinuum: $300M at $5B valuation
IQM: €275M ($320M)
Alice & Bob: €100M Series B
Quantum Machines: $170M Series C

Market projections (McKinsey data presented):

2025: $1.6B market
2030: $7.3B market (34.6% CAGR)
2035: $28B-72B (quantum computing alone)
2040: $198B (total quantum tech market)

My reaction: These are INSANE numbers for a technology that… doesn’t really work yet?

The Technical Breakthroughs (Also Real)

Google’s Willow chip announcement (Dec 2024, discussed at session):

What they achieved:

105 qubits
Exponential error reduction as they scale UP (this is huge)
Tested 3×3 → 5×5 → 7×7 qubit arrays
Error rate cut in half at each step
“Below threshold” — can add qubits WITHOUT adding more errors

The benchmark:

Performed computation in 5 minutes
Would take classical supercomputer: 10 septillion years
(That’s 10,000,000,000,000,000,000,000,000 years)

Why this matters (from Google’s presentation):
Previous quantum computers got MORE error-prone as you added qubits. That’s why they couldn’t scale. Willow reverses this trend.

IBM’s roadmap (from their keynote):

2025: IBM Quantum Loon

New architecture enabling high-rate quantum error correction
Uses qLDPC (quantum low-density parity-check) codes
Slashes error correction overhead by 90%

2029: IBM Quantum Starling

First “large-scale fault-tolerant” quantum computer
200 logical qubits
100 million quantum operations
20,000× more operations than today’s quantum computers

This sounds amazing… except…

The Reality Check (The Part They Don’t Emphasize)

From the “When Will Quantum Actually Matter” panel:

Panelist 1 (Quantum startup CEO):
“We’re making incredible progress. Commercial applications are 3-5 years away.”

Panelist 2 (AWS Quantum, Oskar Painter):
“There’s a tremendous amount of hype in this industry. It can be difficult to filter the optimistic from the completely unrealistic.”

Panelist 3 (MIT Professor, quantum skeptic):
“As of 2025, NO quantum computer has achieved quantum advantage over classical computers in ANY practical task.”

The audience got VERY quiet.

What “quantum advantage” means:
Solving a USEFUL problem faster than classical computers. Not a synthetic benchmark. Not a theoretical speedup. An actual business problem.

Status as of 2025: Zero practical quantum advantages demonstrated.

The NISQ Era Problem

From IBM’s honest assessment:

We’re in the NISQ era:

NISQ = Noisy Intermediate-Scale Quantum
50-1000 qubits (we’re here)
High error rates
Limited scalability
Can’t do error correction (not enough qubits)

What this means in practice:

Quantum computers make mistakes constantly
Every operation has ~0.1-1% error rate
Errors compound exponentially
Results are often garbage

Example from panel:
Running a 100-operation quantum algorithm:

Each operation: 0.5% error rate
By operation 100: Results are 40% wrong
Classical computer: 0% wrong

How do you use a computer that’s wrong 40% of the time?

Answer: You can’t. Not for anything important.

The Error Correction Catch-22

The problem explained (from Google session):

To get useful quantum computing:

Need error correction
Error correction requires 1,000 physical qubits per logical qubit
To run useful algorithms: Need 1,000-10,000 logical qubits
Therefore: Need 1,000,000 to 10,000,000 physical qubits

Current state:

Google Willow: 105 qubits
IBM’s best: 133 qubits
We need: 1,000,000+ qubits

We’re 0.01% of the way there.

IBM’s qLDPC breakthrough:

Reduces overhead from 1,000:1 to ~100:1
Still need 100 physical qubits per logical qubit
Still need 100,000 to 1,000,000 qubits total

We’re 0.1% of the way there.

Timeline for 1M qubit systems: 2035-2040 (maybe)

What Quantum Can’t Do (Despite The Hype)

From the skeptic panelist - debunking common myths:

Myth 1: “Quantum computers will break all encryption”

Reality:

Only breaks RSA and similar algorithms
We already have post-quantum cryptography (classical algorithms quantum can’t break)
Migration to post-quantum crypto is happening NOW
By the time quantum can break RSA, nobody will use RSA

Timeline: This threat is already being mitigated.

Myth 2: “Quantum will revolutionize AI/ML”

Reality:

Most ML is linear algebra and gradient descent
Quantum doesn’t have exponential speedup for these
Classical GPUs are 1000x cheaper
No demonstrated quantum advantage for ML

Quote from AWS panelist: “For AI, quantum is a solution looking for a problem.”

Myth 3: “Quantum will simulate molecules and revolutionize drug discovery”

Reality:

Yes, quantum CAN simulate quantum systems (molecules)
But need fault-tolerant quantum computer (1M qubits)
Classical approximations keep getting better
Drug discovery has other bottlenecks (biology, not simulation)

Timeline: 2030s at earliest, IF 1M qubit systems work.

Myth 4: “Quantum will optimize everything (supply chains, logistics, finance)”

Reality:

Quantum speedup for optimization is quadratic, not exponential
“Quantum only really shines on small-data problems with exponential speedups”
Most real-world optimization: Large data, complex constraints
Classical algorithms + more compute often wins

Quote from panel: “All the rest is beautiful theory, but will not be practical.”

What Quantum MIGHT Actually Be Good For

From the practical applications panel:

1. Cryptography and secure communication (working NOW)

Quantum key distribution (QKD)
Provably secure communication
Already deployed commercially

Status: This actually works today.

2. Quantum sensing

Extremely precise measurements
Magnetic field sensing, navigation, medical imaging
Some commercial products already exist

Status: Real and shipping.

3. Materials science (future)

Simulating quantum materials
Discovering superconductors, batteries, catalysts
Needs fault-tolerant quantum computers

Status: 2030s, maybe.

4. Specific chemistry problems (future)

Nitrogen fixation (fertilizer production)
Catalyst design
Small, specific problems with exponential quantum speedup

Status: 2030s, maybe.

Pattern: The things that work TODAY are NOT computing. The computing applications are 10+ years away.

The Hype Problem

From a16z VC panel - why the hype exists:

Reason 1: It’s genuinely hard

Quantum computing is one of the hardest technical problems humanity has ever attempted.

Requires physics, engineering, computer science, mathematics
Every component must work perfectly
Operating at near absolute zero temperature
Isolating from environment perfectly

Reason 2: Long development cycles

Classical computers: 1940s → 1980s (40 years to mainstream)
Quantum computers: 1980s (theory) → 2020s (early experiments) → 2030s-2040s? (practical)
That’s 50-60 year development cycle

Investors need to believe it’s “almost here” or they won’t fund it.

Reason 3: Competitive pressure

China: $138B government fund for quantum
US: $1.2B National Quantum Initiative
EU: €1B Quantum Flagship program

Governments are in an arms race. Can’t afford to fall behind.

Reason 4: Career incentives

Researchers need to publish exciting results
Startups need to raise money
Vendors need to sell quantum systems
Nobody benefits from saying “this won’t work for 20 years”

Result: Systematic overoptimism.

The “Quantum Winter” Warning

The skeptical MIT professor’s closing statement (everyone was quiet):

"In the 1980s, we had an AI winter. The hype exceeded reality. Funding dried up. Careers ended. It took 30 years to recover.

We’re setting up for a quantum winter.

The promises being made TODAY cannot be delivered in the timeframes being claimed. When that becomes obvious, funding will collapse.

The research will continue, but the startup ecosystem will die.

If you’re investing or building in quantum, ask yourself: Can you survive 10-15 years until this becomes practical?

Because that’s the timeline. Not 3-5 years. Not ‘just around the corner.’ 10-15 years minimum for practical applications.

And maybe longer. Or maybe never, if error correction doesn’t scale as hoped."

The room was SILENT.

Google and IBM presenters did not contradict him.

The Enterprise Reality

From “Quantum for Business” session:

Companies currently “using” quantum:

JPMorgan: Portfolio optimization experiments
Volkswagen: Traffic flow optimization experiments
Pfizer: Molecular simulation experiments

Key word: EXPERIMENTS.

Not production. Not deployed. Experiments.

JPMorgan’s honest assessment (Feb 2025):
“We’ve reduced problem sizes by 80% using quantum-inspired classical algorithms. We haven’t deployed any actual quantum computing in production.”

Translation: Classical algorithms inspired by quantum ideas work better than actual quantum computers.

Companies that shut down quantum teams (2024-2025):

Wells Fargo: Disbanded quantum team
Barclays: Reduced quantum research
Multiple finance companies quietly downscaled

Why:

No path to production in next 5 years
Classical algorithms keep improving
Expensive experiments with no ROI

The Investment Thesis

From the VC panel - why they’re still investing despite skepticism:

Bull case:

Technology is real (physics works)
Progress is happening (error correction improving)
Government support is massive ($138B from China alone)
Whoever wins will dominate (national security implications)
Market could be $200B by 2040

Bear case:

10-15 year timeline minimum
May never achieve practical advantage
Classical computers keep improving (moving target)
Most startups will die in “quantum winter”
Could be another fusion power (always 20 years away)

VC strategy:

Make small bets across multiple companies
Expect 90% to fail
Hope 1-2 become Google/IBM of quantum
10-20 year investment horizon

For founders:

You’re building for 2035-2040, not 2025-2030
Need DEEP pockets or government contracts
Can’t rely on commercial revenue soon
This is a marathon, not a sprint

My Biggest Takeaways

1. The technology is real, but immature

Like computers in 1950. They exist, they work (sort of), but nowhere near practical.

2. Timeline is 10-15 years minimum

Anyone saying “3-5 years to commercial applications” is either lying or delusional.

3. The hype is necessary for funding

Without hype, government and VC funding would dry up. Research would slow.

The hype is part of the strategy.

4. Classical computers keep improving

This is the moving target problem. By the time quantum catches up, classical will be better.

5. Most practical “quantum” applications are actually classical

Quantum-inspired algorithms running on classical computers are beating actual quantum computers.

6. We might be entering “quantum winter”

When promises aren’t met in 3-5 years, funding will crash. Be prepared.

The Uncomfortable Questions

Questions I asked that nobody wanted to answer:

Q1: “What if error correction doesn’t scale as hoped?”

A: Awkward silence. Then: “That would be… problematic.”

Q2: “What if classical algorithms keep improving faster than quantum?”

A: “That’s possible. But we have to try.”

Q3: “How many current quantum startups will still exist in 10 years?”

A: “Probably 10-20%. This is high-risk investment.”

Q4: “Should a regular tech company invest in quantum right now?”

A: “No. Unless you’re a massive enterprise doing R&D for 2035+.”

What I’m Telling My Company

My recommendation to our CEO:

Don’t invest in quantum computing:

Timeline is 10-15 years
No ROI in our investment horizon
Classical solutions work fine

Do watch quantum cryptography:

QKD is real and working
Might be relevant for secure communications
But standard encryption + post-quantum crypto probably fine

Do track progress:

Check in annually on quantum developments
Re-evaluate in 2030 if big breakthroughs happen
Don’t ignore it, but don’t invest yet

Do prepare for post-quantum cryptography:

This is actually urgent
Migrate to quantum-resistant algorithms NOW
NIST standards are finalized

The quantum computing threat to encryption is real.

The quantum computing opportunity for your business is 10+ years away.

For Tech Folks

If you’re considering quantum career:

Pros:

Cutting-edge research
Well-funded (for now)
Intellectually fascinating
Could be next computing revolution

Cons:

Practical applications 10-15 years away
“Quantum winter” risk in next 3-5 years
Most startups will fail
Jobs may disappear if funding dries up

My take: If you’re passionate about physics and okay with long-term research, go for it. If you want to see your work used in production soon, stay in classical computing/AI.

The Honest Timeline

Based on consensus from all sessions:

2025-2027: Continued experiments

Better qubits, lower error rates
More impressive benchmarks on synthetic problems
Zero practical applications

2028-2030: First fault-tolerant systems (maybe)

100-1000 logical qubits
Can run small useful algorithms
First commercial applications in chemistry/materials (maybe)

2031-2035: Scaling up

10,000-100,000 logical qubits
Broader commercial applications
Still expensive, still limited

2036-2040: Mature quantum computing (hopefully)

100,000+ logical qubits
Clear quantum advantages for specific problems
Industry standard for certain applications

Or: Quantum winter happens, progress stalls, and timeline extends to 2040s-2050s.

We don’t know which future we’re in yet.

Bottom Line

Quantum computing in 2025:

Impressive technical progress
Massive investment ($1.25B in Q1)
Huge hype and excitement
Zero practical applications
10-15 year timeline to usefulness
High risk of “quantum winter”

It’s simultaneously real progress AND overhyped.

Both can be true.

Anyone else following quantum? What’s your take: Revolutionary technology being born, or fusion power 2.0 (always 20 years away)?

David

SF Tech Week - Quantum Computing sessions

Sources:

Google Willow quantum chip announcement (Dec 2024)
IBM Quantum Roadmap 2025
McKinsey “Year of Quantum” report (2025)
a16z quantum investment data Q1 2025
Crunchbase quantum funding analysis

security_sam · October 9, 2025, 9:49pm

@product_david - Security engineer here. Let me add the cryptography angle because this is the ONE place where quantum is actually impacting us RIGHT NOW.

My background: Security engineer, responsible for cryptography strategy at fintech company

The “Harvest Now, Decrypt Later” Threat

This is real and happening TODAY:

The attack:

Adversaries (nation states) are recording ALL encrypted traffic NOW
They can’t decrypt it today (RSA, ECC are secure against classical computers)
They store it for 10-20 years
When quantum computers can break RSA/ECC, they decrypt the old traffic
Secrets from 2025 are exposed in 2035

Why this matters:

Medical records (HIPAA violations decades later)
Financial data (trade secrets, M&A communications)
Government secrets (classified data exposed)
Personal communications (blackmail material)

Data retention periods:

Medical: 50+ years
Financial: 10-30 years
Government: Decades to forever

If quantum breaks encryption in 2035, everything encrypted today becomes readable.

What Quantum Actually Breaks

From SF Tech Week “Post-Quantum Cryptography” workshop:

Broken by quantum (Shor’s algorithm):

RSA (used for 80% of internet encryption)
Elliptic Curve Cryptography (ECC)
Diffie-Hellman key exchange
Digital signatures (RSA, ECDSA)

Timeline to break (with fault-tolerant quantum):

RSA-2048: ~8 hours (estimated)
RSA-4096: ~24 hours
ECC-256: ~10 hours

NOT broken by quantum:

AES-256 (symmetric encryption) - quantum only weakens to AES-128 equivalent
SHA-3 (hashing) - quantum-resistant
Lattice-based cryptography - quantum-resistant
Hash-based signatures - quantum-resistant

The NIST Post-Quantum Standards

Major milestone: NIST finalized post-quantum cryptography standards (Aug 2024)

Approved algorithms:

1. CRYSTALS-Kyber (now ML-KEM)

Lattice-based key encapsulation
Replaces RSA/ECC for key exchange
Secure against quantum computers

2. CRYSTALS-Dilithium (now ML-DSA)

Lattice-based digital signatures
Replaces RSA/ECDSA signatures
Quantum-resistant

3. SPHINCS+ (SLH-DSA)

Hash-based signatures
Backup signature scheme
Very conservative, slower but proven secure

These are FINALIZED. Not experimental. Not “maybe.” These are the standards.

The Migration Timeline (It’s Urgent)

From workshop - recommended migration schedule:

2024-2025: Assessment and planning

Inventory all cryptographic systems
Identify what uses RSA/ECC
Prioritize migration (long-lived data first)

2025-2027: Begin migration

Update TLS to support post-quantum algorithms
Hybrid approach (classical + post-quantum)
Start with high-value targets

2027-2030: Complete migration

All systems using post-quantum crypto
Deprecate RSA/ECC for new systems
Maintain backward compatibility

2030+: Post-quantum only

No more RSA/ECC
Fully quantum-resistant infrastructure

We’re in the “you should be doing this NOW” phase.

What We’re Doing (Real Implementation)

Our company’s post-quantum migration (started 2024):

Phase 1 (Completed Q4 2024):

Cryptography inventory (found RSA/ECC in 47 systems)
Risk assessment (prioritized by data sensitivity)
Vendor assessment (do our vendors support post-quantum?)

Phase 2 (In progress 2025):

Updating TLS certificates to hybrid mode (RSA + ML-KEM)
Migrating API authentication to post-quantum signatures
Testing performance impact (post-quantum is slower)

Phase 3 (2026 target):

Full migration of internal systems
Require post-quantum support from all vendors
Deprecate RSA-only systems

Cost so far: $500K (staff time, testing, vendor upgrades)

Timeline: 3 years total

This is REAL work happening NOW. Not a “wait and see” situation.

The Performance Trade-Offs

Post-quantum algorithms are SLOWER and BIGGER:

Key sizes:

RSA-2048: 256 bytes public key
ML-KEM (post-quantum): 1,568 bytes public key
6x larger

Signature sizes:

ECDSA: 64 bytes
ML-DSA: 2,420 bytes
38x larger

Performance:

RSA signature verify: 1,000 ops/sec
ML-DSA verify: 2,500 ops/sec (faster!)
ML-DSA sign: 1,200 ops/sec (slower)

Impact:

TLS handshakes: +100-200ms latency
Certificate chains: +3-5KB size
Mobile bandwidth: Noticeable on slow connections

This is acceptable but not free.

The Quantum Cryptography Misinformation

@product_david mentioned QKD (Quantum Key Distribution). Let me clarify:

What QKD is:

Uses quantum mechanics for provably secure key exchange
Detects eavesdropping (quantum measurement collapses state)
Requires special hardware (quantum photon sources)
Works over fiber optic cable only (100km max distance)

QKD is NOT:

A replacement for encryption (still need AES)
Practical for internet-scale deployment
Cheaper than classical crypto
Necessary for most organizations

QKD is ONLY useful for:

Ultra-high security government/military applications
Very short distances (same datacenter, city-scale)
Situations where you can afford dedicated fiber

For 99.9% of companies: Post-quantum classical crypto is better, cheaper, and more practical.

The Real Quantum Threat Timeline

From workshop - expert consensus:

When will quantum break RSA?

Optimistic (Google/IBM timeline):

2030-2035: First demonstration of breaking RSA-2048
2035-2040: Practical attacks possible

Realistic (skeptic timeline):

2035-2040: Maybe first demonstration
2040-2050: Practical attacks

Pessimistic (might never work):

Error correction doesn’t scale
Quantum computers remain NISQ-era forever
RSA remains secure

Security planning assumption: Plan for 2030-2035 threat.

Why: Because you MUST protect data with 10+ year lifespan NOW.

Even if quantum takes until 2040, data encrypted in 2025 is at risk.

The Vendor Problem

Our biggest challenge in migration:

3rd-party vendors who don’t support post-quantum:

Legacy banking systems (core banking platforms)
Old payment processors
Proprietary HSMs (Hardware Security Modules)
Embedded devices (IoT, point-of-sale terminals)

Many vendors say:
“We’ll support post-quantum when it’s required.”

But:

NIST standards are already finalized
If you wait until it’s “required,” it’s too late
Data encrypted today is vulnerable

We’re putting pressure on vendors:

Requiring post-quantum in new contracts
Penalizing vendors without migration plans
Switching vendors who won’t commit

This is where the real work is: Vendor management, not just our own systems.

What Actually Keeps Me Up At Night

Not quantum computers breaking our encryption in 2035.

What worries me:

1. “Harvest now, decrypt later” for high-value targets

Nation states ARE recording encrypted traffic
They don’t need to decrypt most of it (looking for specific targets)
When quantum works, they’ll have decades of data
We CAN’T protect old data retroactively

2. Slow migration pace

Most companies haven’t started
“We’ll deal with it later” mentality
Later is too late for long-lived data

3. Legacy systems that can’t be upgraded

Medical devices with 20-year lifecycles
Industrial control systems
Banking core systems from 1990s
Can’t just “update the crypto”

4. The hybrid transition period

Need to support both classical and post-quantum
Compatibility issues
Attack surface increases during transition

5. Unknown vulnerabilities in new algorithms

Post-quantum algorithms are NEW
Only ~10 years of analysis
Could have hidden weaknesses
This is why we use hybrid mode (both classical + post-quantum)

The Good News

Unlike quantum computing applications, post-quantum cryptography WORKS TODAY:

Algorithms are finalized
Standards are published
Libraries are available (OpenSSL, BoringSSL support)
Browsers support it (Chrome, Firefox)
Performance is acceptable
Migration is happening

This is not vaporware. This is production-ready.

Recommendations for Security Teams

If you’re responsible for cryptography:

Start NOW:

Inventory all cryptographic systems
Identify RSA/ECC usage
Assess data sensitivity and retention periods
Prioritize: Migrate high-value, long-lived data first

2025-2026:
5. Deploy hybrid TLS (classical + post-quantum)
6. Update internal certificate authorities
7. Migrate APIs to post-quantum signatures
8. Test performance impact

2027-2030:
9. Complete full migration
10. Require vendors support post-quantum
11. Deprecate RSA/ECC

Timeline: Start today. Complete by 2030. Not optional.

For Regular Tech Companies

@product_david asked: “Should regular companies care about quantum?”

For quantum computing: NO. It’s 10-15 years away.

For quantum cryptography: YES. Migration should start NOW.

Minimum actions:

Check if your TLS libraries support post-quantum (most modern ones do)
Enable hybrid mode when available
Ask vendors about their post-quantum migration plans
Budget for cryptography upgrades in next 3 years

Cost: $100K-1M depending on organization size

Timeline: Start in 2025, complete by 2030

This is like Y2K: The impact is real, the timeline is fixed, and delaying makes it worse.

The Irony

The irony of quantum computing:

Quantum computers (the hype):

10-15 years away
Might not work
Overpromised
Mostly irrelevant to businesses

Quantum threat to cryptography (the real impact):

Happening NOW (harvest now, decrypt later)
Definitely works (Shor’s algorithm is proven)
Under-discussed
Requires immediate action

We’re spending billions on quantum computing that might not work.

We’re under-investing in post-quantum cryptography that definitely IS needed.

The priorities are backwards.

The Bottom Line

Quantum computing:

Cool technology
Impressive progress
Overhyped timeline
Not your problem (yet)

Post-quantum cryptography:

Not sexy
Boring migration work
Actually urgent
Definitely your problem

Start migrating to post-quantum crypto TODAY.

Don’t wait for quantum computers to become real. By then it’s too late.

@product_david - you’re right about quantum computing hype. But don’t sleep on the crypto migration. That’s real and urgent.

Sam

SF Tech Week - “Post-Quantum Cryptography” workshop

Sources:

NIST Post-Quantum Cryptography Standards (Aug 2024)
Quantum threat timeline (academic consensus)
Post-quantum migration best practices

cto_michelle · October 9, 2025, 9:49pm

Adding the CTO/investment perspective. I’m on the board of a quantum computing startup (advisor role), so I’ve seen both the promise and the problems up close.

My involvement:

Advisor to quantum computing startup (stealth mode)
Previously CTO at Series C company
Attended a16z quantum VC panel at SF Tech Week

The VC Math on Quantum

From the a16z panel - how VCs think about quantum investing:

Traditional VC math:

5-10 year investment horizon
Expect 30% annual growth
Exit via IPO or acquisition
Portfolio approach (10 bets, 1-2 big wins)

Quantum VC math:

10-20 year investment horizon (!!!)
No revenue for first 5-10 years
Exit via… unclear (who buys quantum companies?)
Most startups will die before technology matures

This is NOT normal VC territory.

Why VCs invest anyway:

1. FOMO (Fear of Missing Out)

“What if this IS the next trillion-dollar industry?”
“What if we miss the next Google?”
Can’t afford to not be in the game

2. Government backstop

Governments will keep funding quantum (national security)
Quantum startups unlikely to completely die (acqui-hire to big tech)
Downside protected somewhat

3. Strategic signaling

Being in quantum makes you look forward-thinking
LP (limited partners) like “deep tech” exposure
Marketing value even if returns are poor

4. Actual belief

Some VCs genuinely believe quantum will work
Physics is sound
Just need engineering scale-up

Quote from one VC: “This is the riskiest bet I’ve made in 20 years. But if it works, it’s a 100x return.”

The Startup Reality

From my quantum startup advisor experience:

Our company (anonymized):

Founded 2021
Raised $40M Series A (2022)
Raised $75M Series B (2024)
Revenue: $2M (mostly research contracts)
Burn: $20M/year
Runway: 5 years

What we do:

Building quantum computer hardware (specific qubit technology)
15 PhDs on team
3 quantum computers built (prototypes)
0 customers

The challenge: We’re building a product for 2035.

How do you run a company for 10+ years with no revenue?

Answer: Government contracts + patient VC money + hope.

The Three Types of Quantum Startups

From a16z panel - quantum startup categories:

Type 1: Full-stack quantum computer companies

Examples: IonQ, Rigetti, PsiQuantum
Building entire quantum computers
Selling access via cloud or direct sales
Extremely capital intensive ($500M+ raised typically)
Long timeline to revenue (5-10 years)
Most will die or get acquired

Type 2: Quantum software/algorithms

Examples: Zapata, QC Ware, Q-CTRL
Building algorithms and software for quantum computers
Selling to enterprises and quantum hardware companies
Less capital intensive ($50-100M raised)
Chicken-and-egg problem: No customers until hardware works
Revenue today: Tiny (mostly consulting/research)

Type 3: Quantum-adjacent

Examples: Quantum Machines (raised $170M), Classiq
Building tools for quantum computers (control systems, software)
Not building qubits themselves
Some revenue today (selling to quantum research labs)
Best near-term prospects

VC consensus: Type 3 has best risk/reward. Type 1 is highest risk/highest reward.

The Revenue Problem

The brutal truth from our startup:

Sources of revenue in quantum (2025):

1. Government research grants

US DOD, DOE, NSF: $500M total/year
EU Quantum Flagship: €1B over 10 years
Competitive, bureaucratic, restricted use

2. Corporate R&D partnerships

JPMorgan, Volkswagen, Pfizer doing quantum experiments
Paying quantum companies $500K-2M/year for access + consulting
Not real deployments, just experiments

3. Cloud access sales

IBM, Amazon, Google sell quantum compute time
~$1-5 per circuit execution
Total market: ~$50-100M/year (tiny)

4. Consulting/services

Teaching enterprises about quantum
“Quantum readiness” assessments
Not scalable, not high-margin

Total addressable market TODAY: ~$1-2B/year

Total quantum startup funding (2024): $1.9B

The industry is raising more money than exists in revenue.

This only works if the market grows 100x in next 10 years.

If it doesn’t: Quantum winter.

What “Quantum Winter” Would Look Like

Scenario (from skeptical VC):

2026-2027: Promises unmet

Quantum companies promised applications by 2025-2027
2027 arrives: Still no practical quantum advantage
Enterprise customers stop experiments (no ROI)

2027-2028: Funding dries up

VCs stop writing checks (burned by unmet promises)
Corporate R&D budgets cut
Quantum startups start dying

2028-2030: Consolidation

70-80% of quantum startups shut down
Big tech (Google, IBM, Amazon) keep research going
Government funding continues but shrinks
Academic research slows

2030+: Recovery?

Technology eventually matures
Practical applications emerge
New wave of startups (but less hype)
OR technology never reaches practical utility

Historical precedent:

AI winter (1980s-2000s): Overpromised, couldn’t deliver, funding crashed
Clean tech bubble (2000s): Billions invested, most companies died
3D printing hype (2010s): Promised revolution, became niche tool

Quantum could follow same pattern.

The Technology Risk

From technical discussions at SF Tech Week:

What could go wrong:

Risk 1: Error correction doesn’t scale

We THINK we can get to 1M qubits with error correction
But what if physical error rates don’t improve enough?
What if overhead stays at 1000:1 instead of 100:1?
Then we need 10M qubits (impossible with current tech)

Risk 2: Decoherence can’t be solved

Quantum states are fragile
Environment interference destroys quantum information
Operating at near absolute zero helps but isn’t enough
What if there’s a fundamental limit we can’t overcome?

Risk 3: Classical algorithms keep improving

This is the moving target problem
Quantum needs exponential speedup to beat classical
But classical computers are getting better too (GPUs, ASICs, better algorithms)
What if classical always wins for practical problems?

Risk 4: Cost never gets reasonable

Current quantum computers: $10-100M each
Require specialized facilities, cooling, shielding
What if costs don’t come down?
Classical supercomputers: $100-200M, but do WAY more

Any of these could doom practical quantum computing.

The China Factor

From geopolitical panel at SF Tech Week:

China quantum investments:

$138B government venture fund (includes quantum)
Micius quantum satellite (launched 2016, operational)
World’s first quantum-encrypted network (1,000km+ fiber)
Leading in quantum communication

US quantum investments:

$1.2B National Quantum Initiative
Private VC: $2-3B/year
Strong in quantum computing research

The competition:

China ahead in quantum communication
US ahead in quantum computing
Both countries view quantum as national security priority
Neither can afford to fall behind

This ensures funding continues even if commercial viability unclear.

Quantum might be like nuclear energy: Strategic asset regardless of economics.

Should Your Company Care About Quantum?

Decision framework from a16z panel:

You should care if:

You’re in finance, pharma, or logistics (potential quantum applications)
You have problems requiring optimization or simulation
You have 10+ year R&D horizon
You have millions to spend on experiments

You should NOT care if:

You’re a typical SaaS/web/mobile company
You need ROI in <5 years
You have limited R&D budget
Your problems are solved fine with classical computing

For 95% of companies: Don’t invest in quantum computing yet.

Exception: DO care about post-quantum cryptography (@security_sam is right).

What I’m Doing

As CTO of a non-quantum company:

Quantum computing:

Watching from sidelines
Annual check-in on progress
Not investing yet
Re-evaluate in 2028-2030

Post-quantum cryptography:

Active migration started 2024
3-year plan to full post-quantum
This is ACTUALLY urgent

As advisor to quantum startup:

Honestly: Worried.

The challenges:

Burn rate is unsustainable without next fundraise
Next fundraise depends on hitting milestones
Milestones keep slipping (physics is hard)
If we miss 2026 fundraise, company dies

We’re racing against both physics AND capital markets.

I give us 40% chance of survival to 2030.

That’s the reality for most quantum startups.

The Honest Assessment

What I told the quantum startup board (real advice):

Plan A: Technology works, we win

Probability: 10-20%
Outcome: Massive success, company worth billions
Timeline: 2030s

Plan B: Technology works, we get acquired

Probability: 20-30%
Outcome: Acqui-hired by Google/IBM/Amazon for team
Timeline: 2027-2030
Return: 2-3x for investors (not great but okay)

Plan C: Technology doesn’t work fast enough, we die

Probability: 40-50%
Outcome: Run out of money before quantum is practical
Timeline: 2026-2028
Return: 0x for investors

Plan D: We pivot to quantum-adjacent business

Probability: 10-20%
Outcome: Become software/tools company for quantum industry
Timeline: 2026-2027
Return: Small outcome but survival

Expected value: Probably negative. But with HUGE upside if it works.

That’s quantum investing.

My Recommendation

For founders: Don’t start a quantum computing company unless:

You’re a quantum physics PhD
You’re okay with 10-20 year timeline
You can raise $50-500M
You’re comfortable with 80% chance of failure

For investors: Only invest in quantum if:

You have 15-20 year horizon
You can afford to lose 100%
You’re making 10+ quantum bets (portfolio approach)
You understand this is deep tech, not typical VC

For CTOs: Don’t invest company resources in quantum computing unless:

You have $10M+ R&D budget
You have specific quantum-amenable problems
You’re okay with zero ROI for 10 years
You have board support for long-term R&D

DO invest in post-quantum cryptography:

Start now
3-5 year migration timeline
Budget $100K-1M
This is not optional

The Uncomfortable Truth

@product_david asked if quantum is real or hype.

Answer: Both.

The physics is real. Quantum computers work. We can manipulate quantum states. Error correction progress is real.

The timeline is hype. “3-5 years to commercial applications” is fantasy. 10-15 years minimum, maybe 20+.

The applications are oversold. Quantum won’t revolutionize everything. It will be useful for specific problems. Most companies won’t need it.

The investment is rational (sort of). High risk, high reward, long timeline. VCs know this. They’re making calculated bets.

Most startups will die. This is not controversial. Everyone knows it. It’s priced in.

But the winners could be HUGE. If Google or IBM crack this, it’s a trillion-dollar industry.

So the money keeps flowing.

My Prediction

10 years from now (2035):

Scenario A (30% probability): Quantum works

Fault-tolerant quantum computers exist
100,000+ logical qubits
Commercial applications in chemistry, materials, optimization
Market is $50B+ annually
3-5 quantum computing giants (Google, IBM, + 1-2 startups)
We remember this as the birth of quantum computing era

Scenario B (50% probability): Quantum partially works

Quantum computers improve but remain limited
1,000-10,000 logical qubits
Niche applications only
Market is $5-10B annually
Only Google, IBM, Amazon still investing heavily
Viewed as expensive research tool, not revolution

Scenario C (20% probability): Quantum doesn’t work

Error correction hits fundamental limits
Can’t scale beyond NISQ-era
No practical quantum advantage achieved
Market collapses to <$1B (mostly research)
Viewed as failed technology (like fusion power)
We remember this as “quantum winter”

My bet: Scenario B.

Quantum will work, but not as well or as broadly as promised.

It will be a powerful specialized tool, not a universal revolution.

And that’s okay. Not everything has to change the world.

Michelle

SF Tech Week - a16z quantum VC panel + personal experience

Sources:

a16z quantum investment data
Personal experience as quantum startup advisor
VC panel discussions SF Tech Week

product_david · October 9, 2025, 9:49pm

@security_sam @cto_michelle - This thread perfectly captures the quantum paradox: Real technology, absurd hype, and one actually urgent problem (crypto migration).

What I’m Taking Away

On quantum computing:

@cto_michelle’s Scenario B (50% probability) feels right:

Quantum computers will improve
They’ll become useful for SPECIFIC problems
But not the universal revolution being promised
Timeline: 2030s at earliest

Key insight from @security_sam:

The ironic inversion:

Everyone hyped about quantum computing (10+ years away)
Nobody talking about post-quantum crypto (urgent NOW)
Priorities are exactly backwards

This is like worrying about Mars colonization while ignoring climate change.

What Changed My Mind

Before SF Tech Week, I thought:
“Quantum is total vaporware. It’ll never work.”

After SF Tech Week, I think:
“Quantum is real physics making real progress toward practical applications in 10-15 years, but the 3-5 year hype is nonsense, most startups will die, and the only urgent problem is crypto migration.”

That’s a much more nuanced view.

The Pattern I’m Seeing

This reminds me of AI trajectory:

1980s AI: “AI will be everywhere in 10 years!”
1990s: AI winter, funding crashed, startups died
2000s: Slow steady progress, less hype
2010s: Deep learning breakthrough
2020s: AI actually works, becomes mainstream

Timeline: 40 years from promise to delivery.

Quantum computing might follow same path:

2020s: “Quantum will be everywhere in 10 years!”
2030s: Quantum winter? Or slow steady progress
2040s: Breakthroughs?
2050s: Actually practical?

Timeline: 30-40 years from promise to delivery.

We’re in the hype phase. The “AI winter” equivalent is coming.

What I’m Telling My CEO (Updated)

Original recommendation: Ignore quantum entirely.

Updated recommendation after this thread:

Quantum computing:

Ignore for product development (10-15 year timeline)
Monitor annually for breakthroughs
Re-evaluate in 2030 if fault-tolerant systems emerge

Post-quantum cryptography (@security_sam convinced me):

Start migration planning NOW
Budget $200K for crypto assessment + initial migration
3-year timeline to full post-quantum
This is ACTUALLY urgent

Quantum sensing/communication:

Not relevant for our business (fintech)
Might matter for specific industries (medical, defense)

The Uncomfortable Questions (Still Unanswered)

Nobody at SF Tech Week could answer these:

Q1: What if classical computers improve faster than quantum computers?

This is the moving target problem. By the time quantum can beat classical in 2035, what will classical computers look like?

A: Awkward silence. “We’re betting quantum speedup is exponential enough to stay ahead.”

Q2: What if error correction doesn’t scale as hoped?

We THINK we can get to 1M qubits. But what if we can’t?

A: “That would be… problematic.” (Nobody wants to think about this.)

Q3: How many quantum startups will still exist in 2030?

A: “10-20% probably.” (80-90% will die.)

Q4: Is this the next fusion power? (Always 20 years away)

A: Nervous laughter. No real answer.

These questions make people uncomfortable because they’re real risks.

The One Thing That’s Actually Urgent

@security_sam’s warning about “harvest now, decrypt later”:

This changed my thinking.

The threat is:

Nation states record encrypted traffic TODAY
They can’t decrypt it yet
In 2035, when quantum works, they decrypt everything from 2025
Your secrets from 2025 are exposed in 2035

For data with >10 year lifespan, this is REAL.

Medical records, financial data, trade secrets, personal communications.

Even if quantum computing never becomes practical, the THREAT to cryptography is enough to require action now.

My Action Items

Personal:

Keep watching quantum (annual check-in)
Don’t invest in quantum startups (too risky)
Curious to see if @cto_michelle’s quantum startup survives (40% chance she said)

Professional:

Push our CTO to start post-quantum crypto assessment
Budget $200K for initial migration
Timeline: 3 years to complete

Community:

Share this thread with other product/engineering leaders
Help correct the “quantum will save us in 3-5 years” misconception
Spread awareness of actual urgent problem (crypto migration)

The Honest Timeline (Synthesis)

Combining all perspectives:

2025-2027: Hype continues

Google/IBM/Amazon make impressive announcements
More benchmarks on synthetic problems
VCs keep investing ($1-2B/year)
Zero practical applications

2027-2030: Reality check

Promised applications don’t materialize
Enterprise customers stop experiments
Some quantum startups die
Possible “quantum winter” funding crunch

2030-2035: Slow progress or breakthrough?

Fork in road: Does fault-tolerant quantum work?
IF YES: First commercial applications emerge
IF NO: Quantum relegated to research tool

2035-2040: Maturity or collapse

Either quantum becomes practical tool for specific problems
Or technology stalls and funding dries up permanently

My bet: Scenario B (useful for specific problems, not universal revolution).

What I Learned at SF Tech Week

The good:

Quantum physics is real and progress is measurable
Error correction is improving (Google Willow breakthrough)
Investment is massive ($1.25B in Q1 2025)
Some very smart people genuinely believe this will work

The bad:

Timeline is 10-15 years minimum, not 3-5 years
Most startups will die before technology matures
Classical computers keep improving (moving target)
“Quantum winter” risk is very real

The urgent:

Post-quantum cryptography migration must start NOW
“Harvest now, decrypt later” threat is real
NIST standards are finalized and ready
3-5 year migration timeline

The irony:

We’re hyping the 10-15 year opportunity
We’re ignoring the immediate crisis
Classic human behavior

For Other Tech Leaders

Key takeaways:

Don’t invest in quantum computing yet (unless you’re massive enterprise with 10+ year R&D horizon)
DO start post-quantum crypto migration (this is not optional, start NOW)
Watch quantum progress annually (re-evaluate in 2028-2030)
Be skeptical of “3-5 year” timelines (real timeline is 10-15+ years)
Most quantum startups will fail (80-90% based on VC estimates)
But winners could be huge (if it works, it’s trillion-dollar industry)

Thanks @security_sam for crypto wake-up call and @cto_michelle for insider VC/startup perspective. This thread is going in my “hype vs. reality” reference file.

Anyone else attending quantum sessions at SF Tech Week? What’s your take?

David

SF Tech Week - synthesis of quantum discussions