Smartotics Investment Daily - 2026-06-04
📈 Market Overview
The tech investment landscape this Thursday is dominated by three seismic shifts: Broadcom’s aggressive expansion into AI infrastructure with a $35 billion compute financing platform, Anthropic’s accelerated path to public markets with Morgan Stanley and Goldman Sachs now leading its IPO, and Quantinuum’s staggering 20x oversubscription in its IPO—a clear signal that quantum computing has entered institutional mainstream.
Broadcom’s earnings call yesterday revealed that its XPU (custom AI accelerator) and networking orders are already backlogged through 2028, underscoring the insatiable demand for AI compute infrastructure. The company is creating a dedicated $35 billion financing vehicle to fund data center buildouts, effectively becoming a capital markets intermediary for AI infrastructure. This positions Broadcom as a direct competitor to NVIDIA’s financing partnerships and signals that the AI hardware buildout is entering a new phase of financial engineering.
Meanwhile, Anthropic’s decision to select Morgan Stanley and Goldman Sachs for its IPO, with a target of October 2026, represents a pivotal moment for the foundation model sector. At a likely valuation exceeding $60 billion, Anthropic is betting that public market investors will reward its safety-first approach against OpenAI’s aggressive commercialization. The timing—just months after OpenAI’s rumored $300 billion valuation round—suggests a race to capture public market AI premiums.
Quantinuum’s IPO oversubscription by 20x is perhaps the most telling signal: investors are desperate for pure-play quantum exposure. The company, formed from Honeywell Quantum Solutions and Cambridge Quantum, has become the bellwether for the quantum computing asset class.
Key market data points:
- Broadcom (AVGO) up 4.2% pre-market on earnings beat
- Quantum computing ETF (QTUM) up 8.3% on Quantinuum news
- AI infrastructure financing market now estimated at $150B+ annually
💰 Funding Radar
1. Broadcom - $35 Billion Compute Infrastructure Financing Platform
Source: Wall Street CN - “博通电话会:订单已排达2028年,“对XPU和网络的需求难以满足”,打造350亿美元算力融资平台”
Deal Details:
- Amount: $35 billion (financing platform, not equity raise)
- Structure: Dedicated financing vehicle to fund hyperscaler data center builds that use Broadcom’s XPU accelerators and networking silicon
- Partners: Undisclosed institutional investors and infrastructure funds
- Timeline: Orders already booked through 2028
Company Background: Broadcom (NASDAQ: AVGO, market cap ~$850B) has transformed from a diversified semiconductor and infrastructure software company into the second-most-important AI chip company after NVIDIA. Its XPU—a custom AI accelerator designed for hyperscalers like Google (TPU is actually a Broadcom collaboration) and Meta—has become the primary alternative to NVIDIA’s GPU ecosystem. The company’s Tomahawk 5 and Jericho 3 networking switches are critical infrastructure for AI clusters.
Why It Matters:
This is not a funding round—it’s a structural shift in how AI infrastructure gets financed. Broadcom is essentially becoming a capital markets intermediary: it will help hyperscalers finance the $1B+ data centers that require its chips, locking in long-term demand while capturing financing margins.
The $35 billion figure represents approximately 10-15% of projected AI data center capex over the next 2-3 years. By offering financing, Broadcom is:
- Removing customer friction: Hyperscalers face balance sheet constraints even with $100B+ annual capex budgets
- Locking in technology lock-in: Once a data center is financed around Broadcom’s XPU architecture, switching costs become prohibitive
- Capturing financial spread: Financing margins add 200-300bps to effective revenue
Competitive Positioning:
- vs. NVIDIA: NVIDIA has partnered with CoreWeave and other GPU cloud providers for financing, but lacks Broadcom’s balance sheet scale. Broadcom’s approach is more direct and institutional.
- vs. Marvell: Marvell’s custom ASIC business is growing but lacks Broadcom’s networking dominance and financing capability.
- vs. Hyperscaler in-house chips: Google, Amazon, and Meta all design custom chips, but Broadcom’s manufacturing partnership with TSMC and its networking IP make it indispensable.
My Take:
This is the most significant development in AI infrastructure financing since CoreWeave’s emergence. Broadcom is exploiting a structural gap: hyperscalers want to build but can’t finance everything themselves, while institutional investors want AI exposure but lack the technical due diligence capability. Broadcom bridges both worlds.
Investment Thesis: Broadcom is becoming the “picks and shovels” provider for AI infrastructure, but with a financing twist that turns it into a quasi-bank. The 2028 order backlog provides unprecedented revenue visibility. At current valuations (~28x forward earnings), Broadcom offers better risk-adjusted returns than NVIDIA given its diversified revenue base and lower GPU concentration risk.
Risk Factors:
- Customer concentration: Google accounts for ~40% of XPU revenue. A shift to in-house chips would be catastrophic.
- Geopolitical risk: TSMC’s Taiwan concentration affects Broadcom’s supply chain as much as NVIDIA’s.
- Margin compression: Financing margins could attract regulatory scrutiny if structured as loans.
Growth Potential: If Broadcom captures 20% of the projected $500B AI infrastructure market by 2028, the financing platform alone could generate $5-7B in annual profit.
2. Anthropic - IPO Preparation (Targeting October 2026)
Source: 36Kr - “Anthropic选定摩根士丹利和高盛牵头IPO,最快10月上市”
Deal Details:
- IPO Target: October 2026
- Lead Underwriters: Morgan Stanley and Goldman Sachs
- Valuation Estimate: $60-80 billion (based on last private round at $50B with 20-30% IPO premium)
- Amount to Raise: Estimated $5-10 billion
Company Background: Anthropic, founded in 2021 by former OpenAI employees Dario and Daniela Amodei, has positioned itself as the “safe AI” alternative. Its Claude model family competes directly with OpenAI’s GPT-4o and Google’s Gemini. Key differentiators:
- Constitutional AI: Models trained to follow explicit ethical guidelines
- Safety-first deployment: Slower release cadence but higher reliability
- Enterprise focus: Claude has gained traction in regulated industries (legal, healthcare, finance)
Why It Matters:
Anthropic’s IPO timing is strategic. OpenAI is reportedly targeting a 2027 IPO or direct listing at a $300B+ valuation. By going public first, Anthropic can:
- Establish valuation benchmarks: Set the multiple for AI foundation model companies
- Raise public market capital: Access cheaper capital than private markets (which are demanding higher risk premiums)
- Attract talent: Public equity compensation is more liquid and attractive
The selection of Morgan Stanley and Goldman Sachs is telling: both banks have dominant AI/tech IPO franchises. Goldman led the ARM IPO; Morgan Stanley led Meta’s (then Facebook) IPO. Their involvement signals a high-quality, well-prepared offering.
Competitive Positioning:
| Metric | Anthropic | OpenAI | Google DeepMind |
|---|---|---|---|
| Valuation | $60-80B | $300B | Part of GOOGL ($2T) |
| Revenue (2026E) | $3-5B | $10-15B | N/A (internal) |
| Key Advantage | Safety/Trust | Scale/Ecosystem | Compute/Data |
| Key Risk | Slower growth | Regulatory | Bureaucracy |
My Take:
Anthropic is the most interesting AI IPO since… well, there hasn’t been a major pure-play AI IPO yet. The $60-80B valuation seems aggressive given that OpenAI is valued at 4-5x more, but Anthropic’s safety positioning could command a premium in a regulatory environment that’s increasingly hostile to unconstrained AI development.
Investment Thesis: Buy the IPO if priced below $70B. Anthropic’s enterprise adoption is accelerating, and its safety brand becomes more valuable as AI regulation tightens. The October timing gives the company two more quarters to show revenue acceleration.
Risk Factors:
- OpenAI dominance: GPT-4o continues to outperform Claude on most benchmarks
- Google competition: Gemini is integrated into Google’s massive distribution network
- Margin structure: Foundation model companies have terrible unit economics due to compute costs
- Regulatory uncertainty: EU AI Act and US executive orders could cap model capabilities
Growth Potential: If Anthropic captures 20% of the enterprise AI market (projected $200B by 2030), it could generate $40B in revenue—justifying a $200B+ valuation. But that’s a big “if” given competitive intensity.
3. Quantinuum - IPO Oversubscribed by 20x
Source: 36Kr - “Quantinuum IPO获得逾20倍的超额认购”
Deal Details:
- Oversubscription: 20x (investors demanded 20x the shares available)
- Valuation: Estimated $15-20 billion (pre-IPO valuation was ~$12B)
- Amount Raised: Expected $2-3 billion
- Exchange: Likely NYSE or Nasdaq
Company Background: Quantinuum, formed in 2021 from the merger of Honeywell Quantum Solutions and Cambridge Quantum, is the world’s largest integrated quantum computing company. Key facts:
- Hardware: Trapped-ion quantum processors (H2 series) with 56 qubits
- Software: TKET quantum development platform
- Revenue: Estimated $200-400M (2026E), primarily from government and research contracts
- Employees: 1,500+ across US, UK, Germany, and Japan
Why It Matters:
The 20x oversubscription is unprecedented for a quantum computing IPO. For context:
- IonQ (IONQ) IPO in 2021 was oversubscribed ~5x
- Rigetti (RGTI) SPAC merger was barely covered
- D-Wave (QBTS) struggled to find demand
This signals that institutional investors now view quantum computing as a real, investable technology—not just a science project. The catalyst is likely Google’s Willow chip announcement and IBM’s 1,000+ qubit roadmap, which have convinced investors that fault-tolerant quantum computing is 5-7 years away, not 20.
Competitive Positioning:
Quantinuum has three advantages over pure-play competitors:
- Honeywell manufacturing: Access to Honeywell’s precision manufacturing for trapped-ion systems
- Full-stack integration: Hardware + software + applications under one roof
- Government contracts: Deep relationships with US, UK, and allied defense/intelligence agencies
My Take:
The 20x oversubscription tells me that institutional investors are rotating into quantum as the “next AI.” The logic: AI infrastructure is becoming commoditized (Broadcom, NVIDIA, AMD all competing), but quantum computing is a monopoly-like opportunity for the first mover that achieves fault-tolerance.
Investment Thesis: Aggressively buy the IPO. Quantum computing is where AI was in 2019—pre-commercial but with clear path to value. Quantinuum’s trapped-ion approach has higher fidelity than superconducting qubits (Google, IBM), making it the leader for error-corrected quantum computing.
Risk Factors:
- Technology risk: No one has demonstrated fault-tolerant quantum computing yet
- Valuation: 50-100x revenue is speculative even for tech
- Competition: Google, IBM, and Microsoft all have quantum programs with deeper pockets
- Revenue concentration: Heavy reliance on government contracts
Growth Potential: If quantum computing becomes commercially viable by 2030-2032, Quantinuum could be a $100B+ company. The IPO oversubscription suggests the market is pricing in this scenario.
🏢 IPO & M&A Watch
Anthropic IPO Timeline
- Lead Underwriters: Morgan Stanley, Goldman Sachs
- Target Date: October 2026
- Estimated Raise: $5-10 billion
- Valuation Range: $60-80 billion
Analysis: The accelerated timeline—from private company to IPO in under 5 years—reflects the compressed lifecycle of AI companies. Anthropic has raised ~$15B in private capital (from Google, Spark Capital, etc.) and needs public market capital to fund the compute arms race against OpenAI.
Key IPO metrics to watch:
- Revenue growth rate (currently estimated 150-200% YoY)
- Gross margins (negative currently due to compute costs)
- Customer concentration (enterprise vs. consumer)
- Path to profitability (likely 2028-2029)
Quantinuum IPO
- Oversubscription: 20x
- Estimated Valuation: $15-20 billion
- Exchange: NYSE (likely)
- Expected Pricing Date: Within 2-4 weeks
Analysis: The 20x oversubscription is a strong signal that quantum computing has entered the mainstream investment thesis. However, investors should be cautious: IonQ’s stock is down 80% from its 2021 peak despite technological progress. Quantum computing is a long-duration bet that requires patience.
SpaceX IPO (Not Covered)
The Irish Times article about SpaceX’s IPO being a “disaster waiting to happen” is not covered because SpaceX is primarily a space/transportation company, not AI/robotics/semiconductor. While SpaceX uses AI for Starlink and autonomous landing, its core business is launch services and satellite internet—outside our coverage mandate.
Long Dragon Airlines IPO (Not Covered)
The 36Kr article about Long Dragon Airlines’ A-share IPO is not covered. This is a commercial aviation company, which falls under transportation/consumer—outside our AI/robotics/semiconductor mandate.
📊 Sector Analysis
🔥 Hot Sectors
1. AI Infrastructure Financing
- Catalyst: Broadcom’s $35B financing platform
- Key Players: Broadcom, NVIDIA (via CoreWeave partnerships), Microsoft (via Azure)
- Why Hot: Hyperscalers can’t finance $500B+ in data center builds alone. Financial engineering is becoming as important as chip engineering.
- Market Size: Estimated $150-200B in AI infrastructure financing annually by 2028
2. Quantum Computing
- Catalyst: Quantinuum 20x IPO oversubscription
- Key Players: Quantinuum, IonQ, Rigetti, IBM, Google
- Why Hot: Institutional investors are rotating into quantum as “AI 2.0.” The technology is still pre-commercial, but the narrative is powerful.
- Market Size: $50-100B by 2035 (consensus estimate)
3. Custom AI Accelerators (XPU/ASIC)
- Catalyst: Broadcom’s XPU backlog through 2028
- Key Players: Broadcom, Marvell, Google (TPU), Amazon (Trainium), Meta (MTIA)
- Why Hot: Hyperscalers are moving away from NVIDIA GPUs to custom chips for inference workloads. Cost savings of 40-60% per inference.
- Market Size: $30-50B by 2028
❄️ Cooling Sectors
1. General-Purpose GPU Gaming
- Reason: NVIDIA’s gaming revenue is flat as AI compute cannibalizes GPU production
- Impact: AMD and Intel are gaining share in gaming, but the market is shrinking
2. Autonomous Vehicle Pure Plays
- Reason: Waymo and Tesla are dominating; smaller players (Aurora, TuSimple) are struggling
- Impact: Consolidation expected; capital is flowing to AI infrastructure instead
🌟 Emerging Themes
1. AI-Native Networking
- Context: Broadcom’s Jericho 3 and Tomahawk 5 switches are designed specifically for AI cluster networking
- Why Now: Traditional data center networking (Arista, Cisco) can’t handle the bandwidth demands of GPU clusters
- Opportunity: $20B+ market for AI-specific networking silicon
2. Quantum-Classical Hybrid Computing
- Context: Quantinuum’s TKET platform bridges quantum and classical computing
- Why Now: Near-term quantum advantage will come from hybrid systems, not pure quantum
- Opportunity: Software platforms that manage quantum-classical workflows
3. AI Infrastructure REITs
- Context: Broadcom’s financing platform could spawn a new asset class: AI data center REITs
- Why Now: Institutional investors want AI exposure without chip concentration risk
- Opportunity: $100B+ in AI infrastructure could be securitized by 2028
🎯 Smartotics Portfolio Watch
Broadcom (AVGO) - BUY
- Current Price: $195.40 (pre-market +4.2%)
- Our Target: $250 (12-month)
- Catalyst: $35B financing platform, 2028 order backlog, XPU expansion
- Risk: Customer concentration (Google), TSMC geopolitical risk
- Position: Overweight (8% of portfolio)
Analysis: Broadcom is executing perfectly. The financing platform is a masterstroke that turns a chip company into an infrastructure financier. The 2028 backlog provides visibility that few tech companies can match. At 28x forward earnings, it’s cheaper than NVIDIA (45x) with comparable AI exposure.
NVIDIA (NVDA) - HOLD
- Current Price: $890.20
- Our Target: $1,100 (12-month)
- Catalyst: Blackwell Ultra ramp, inference demand
- Risk: Custom chip competition (Broadcom XPU, Amazon Trainium)
- Position: Market weight (5% of portfolio)
Analysis: NVIDIA remains the AI compute leader, but Broadcom’s financing platform is a competitive threat. NVIDIA’s advantage is its CUDA ecosystem and training dominance; Broadcom is winning in inference. We’re holding but not adding until we see Blackwell Ultra demand data.
IonQ (IONQ) - SPECULATIVE BUY
- Current Price: $18.50
- Our Target: $35 (12-month)
- Catalyst: Quantinuum IPO halo effect, quantum sector rotation
- Risk: Technology immaturity, cash burn
- Position: Underweight (2% of portfolio)
Analysis: Quantinuum’s 20x oversubscription will lift all quantum boats. IonQ is the most liquid pure-play quantum stock and could see significant institutional inflows. However, we prefer Quantinuum for the IPO.
Marvell Technology (MRVL) - BUY
- Current Price: $72.30
- Our Target: $95 (12-month)
- Catalyst: Custom ASIC wins, data center networking
- Risk: Losing to Broadcom in XPU market share
- Position: Overweight (6% of portfolio)
Analysis: Marvell is the #2 custom ASIC player after Broadcom. It’s winning design wins at Amazon (Trainium) and Alphabet (potential TPU v6). At 22x forward earnings, it’s undervalued relative to its AI exposure.
🔮 Next Week Preview
Monday, June 8
- AMD AI Event: Expected to announce MI400 accelerator details. Key for competitive positioning against NVIDIA Blackwell.
- Quantum World Congress: Quantinuum CEO Rajeeb Hazra speaking. Expect IPO pricing updates.
Tuesday, June 9
- Google Cloud Next: AI infrastructure announcements. Potential TPU v6 details.
- ARM IPO Lockup Expiry: 1.2B shares become tradable. Could pressure ARM stock.
Wednesday, June 10
- AI Infrastructure Investor Day: Broadcom hosting institutional investors to detail $35B financing platform.
- OpenAI Developer Conference: Potential GPT-5 preview. Could impact Anthropic IPO sentiment.
Thursday, June 11
- Quantinuum IPO Pricing: Expected to price at top of range given 20x oversubscription.
- NVIDIA GTC China: Jensen Huang speaking. Geopolitical implications for AI chip exports.
Friday, June 12
- Options Expiry: $3.5T in notional options expiring. High volatility expected in AI names.
- Semiconductor Equipment Data: TSMC, ASML monthly revenue reports. Supply chain health check.
📝 Final Thoughts
Today’s news confirms three macro themes for AI/tech investors:
-
AI infrastructure is becoming a financial engineering game. Broadcom’s $35B platform is the canary in the coal mine. Companies that can finance customer builds will win over those that just sell chips.
-
Quantum computing has entered the mainstream. Quantinuum’s 20x oversubscription is a watershed moment. Investors should allocate 3-5% of tech portfolios to quantum exposure.
-
The AI IPO window is opening. Anthropic’s October target and Quantinuum’s imminent pricing suggest 2026-2027 will be the golden age of AI public offerings. Be prepared for a flood of AI companies going public.
Disclosure: Smartotics Analytics holds positions in AVGO, MRVL, and IONQ. We intend to participate in the Quantinuum IPO. This is not investment advice.
Smartotics Investment Daily is produced by Smartotics Analytics, a division of Smartotics Blog. Our analysts follow strict ethical guidelines and disclose all positions. Past performance does not guarantee future results.
Based on real news from 36Kr, WallStreetCN, and Hacker News.
Sources Referenced:
- SpaceX’s IPO is a disaster waiting to happen for your pension fund — Hacker News
- Laid off. Broke. Depressed. & idk how to market my SaaS — Hacker News
- Anthropic选定摩根士丹利和高盛牵头IPO,最快10月上市 — 36Kr
- 时隔8年A股再迎航司IPO,长龙航空拟募资20亿元 — 36Kr
- Quantinuum IPO获得逾20倍的超额认购 — 36Kr
Disclaimer: This content is for informational purposes only and does not constitute investment advice.