We're excited to announce that two popular FUNDA Plays, GPU Rental Price Tracker and Thematic Sector Rotation, are now available to all our followers, with no login required. GPU Rental Price Tracker Compare GPU rental prices across providers and lease types, including spot, on-demand, reserved, and capacity block pricing. The dashboard brings cross-vendor price indices and historical trends together to provide a consistent reference for monitoring changes in compute costs. Start here: funda.ai/play/gpu-spot-price… Thematic Sector Rotation Track relative strength and momentum across thematic sectors, compare daily rankings, and monitor shifts in market leadership. The dashboard brings rotation charts, leadership maps, and market-style indicators together to help assess how sector trends are evolving. Start here: funda.ai/play/sector-rotatio… If you find these Plays valuable, subscribe to FUNDA to unlock the full capabilities of our research platform.

$META 3Q26 revenue: $64.60bn. Our preview: $64.66bn (-0.1%). Street: $63.35bn. Why we were above consensus: ad agency checks raised Q3 Meta budgets, plus WhatsApp Business messaging, Audience Network, travel, midterm election ads & a smaller FX drag. FUNDA Report funda.ai/reports/preview-met… [email protected]

Akram, sharp as ever. @akramsrazor

The PHP2Shell vulnerability is an important signal: AI is significantly lowering the barrier to discovering and analyzing sophisticated vulnerabilities. Over the next 12–24 months, we could see a wave of long-standing vulnerabilities that were previously too difficult to uncover. This suggests that AI could create incremental demand for cybersecurity, rather than simply shifting existing spending. AI labs excel at vulnerability discovery, while security platforms like $NET, $PANW, and $crwd specialize in detecting, preventing, and responding to threats in real-world environments. Over the short to medium term, we see AI labs and security platforms as complementary rather than competitive.

Preview| $NOW 3Q26: Solid Checks and a Healthy 2027 Outlook, AICT Adoption Starts to Take Shape Checks point to a modest 3Q beat, with renewals and AI tier upgrades supporting stronger 2027 growth. We revisited three ServiceNow channel experts this quarter and spoke with customers in insurance and healthcare. Feedback remains solid, with two channels modestly ahead of their 3Q plans and the third tracking within expectations. Renewals, AI tier upgrades and module expansion support a healthy 4Q and 2027 outlook, extending our view from last quarter. On AI, AICT is moving beyond discussions into discovery, governance workflows and live-environment pilots. The strategic opportunity is becoming more tangible and could become a more meaningful revenue driver in 2027, despite its limited near-term contribution. NOW is not necessarily the outright governance winner, but success with AICT could make its role in the AI era more critical. Now Assist sales and deployment are improving; incremental consumption and customer ROI remain more uneven. We expect a 150-200bps beat on 3Q cRPO growth is reasonable, while 4Q cRPO growth guidance is likely to be in line with buyside expectations at around 20%. Detailed Report fundaai.substack.com/p/previ…

In our OCP note, we discuss the $NVDA critique of NPO, the impact of micro-VCSEL on $lite, the impact of XPO and USFP on NPO and CPO, and constraints in the shape of light sources, fiber and testing. fundaai.substack.com/p/resea…

Research|Power and Liquid Cooling: OCP 2026 Sharpens Power Deployment Timelines; Single-Phase Cooling Has Room to Improve OCP showed earlier 800V timing and a longer single-phase cooling runway, but no major architecture shift. 1. In power and liquid cooling, there were no major architectural changes at OCP this year; the movement is in specs, timing and detail. 2. The NVIDIA and AMD keynotes restated existing roadmaps, and AMD’s only new launch, the Helion agentic rack, is a product-line extension rather than an architectural change. 3. The break-out sessions did show finer changes: NVIDIA wants SST deployment readiness by 2028, earlier than investors generally expect; the runway for single-phase cold plates remains intact and should stay mainstream through 2030; capacitors inside the PSU may squeeze out the CBU before it ever scales; and the ±400 V versus 800 V argument has converged on voltage-to-ground, so tray-side specs are compatible and suppliers are not facing two product lines. 4. We attended the 2026 OCP Global Summit in San Jose from October 12 to 15. Our note on power and liquid cooling below draws on learnings from the show floor and 95 sessions, 22 of which we attended in person. 5. We identify three key themes in power and liquid cooling and discuss the debates surrounding each theme, before ending the note with a roundup of discussions with investors as well as some photographs from the show floor. Detailed Report fundaai.substack.com/p/resea…

We are forecasting $SPCX 26Q3 revenue of $13.0b today and are sharing a free dashboard that we hope you will find useful. funda.ai/play/spacex?ref=cat…

Deep|LLM: Enterprise AI Adoption Survey (Vol. 6) - In Our Sample: AI Spending Growth Expected to Slow; Routine Work Shifts to Lower-Cost Models AI spending growth is expected to slow at most companies in this sample. The fintech group’s spending growth is expected to slow after a rapid initial AI rollout. At the pharmaceutical group, better model selection and procurement terms are expected to curb spending growth. Spending growth at the pharmacy and health insurance group is expected to remain broadly steady. Claude-based workflows can replace specialist software for feature engineering. The subprime auto lender’s in-house workflow lets its team direct the search for variables, require explanations of their business meaning and account for compliance constraints. With that workflow meeting its needs, the lender plans to cancel or not renew a specialist tool costing ~$80K-100K a year. Matching models to tasks can cut AI spending even as token use rises. After training, some teams at the fintech group roughly halved AI spending while using more tokens. The pharmacy and health insurance group’s expert estimates that better routing and model choice could reduce API costs by 20%-30% for the same workload; its automatic router is still under evaluation. Some companies tie additional AI funding to expected cost savings, revenue gains or less manual work. The brewing group asks business owners to explain expected cost savings or additional revenue and whether they will commit to changing workflows and reducing headcount after full implementation. At the pharmacy and health insurance group, proposals to scale a use case must explain how it will reduce manual processing, handling time or rework. Most companies in this sample have not adopted Fable for routine production; one interviewed company reports using Astra for cross-system commercial analysis. Cost, a lack of recurring tasks that justify it, and data-retention terms constrain broader Fable adoption. The mega brewing group uses Astra to combine market strategy, brand plans and financial information with customer-pricing information. Self-hosted open-weight models already handle about half of API calls at one interviewed company. The pharmaceutical group has incorporated model downloading, scanning, verification and tuning into its internal process. Its expert roughly estimates model-use costs to be ~90% lower than with external APIs. These workloads avoid external per-token charges but still incur internal deployment and operating costs. Detailed Report fundaai.substack.com/p/deepl…

A few takeaways on ChatGPT advertising and what it means for Meta, based on FUNDA’s latest META preview and checks with 6 ad agencies: 1/ ChatGPT ads are growing quickly, but scale and budget impact are two different questions. According to the report, ChatGPT’s ad revenue run-rate reached $1bn by August 31, following the start of testing in February. Ads were live in more than 60 countries by September 23. The agency checks show adoption is rising, though from a small base: A. North America Agency A: ~55–58 advertisers are testing ChatGPT ads, but only ~7–8 have moved into regular budgets. ChatGPT still represents less than 1% of the total digital ad budget the agency tracks. B. Asia Agency A: ChatGPT spend rose from ~1/60 of Google spend in 2Q to ~1/10 in 3Q. The agency expects ~1/8–1/7 in 4Q. C. Asia Agency C: ~170 advertisers are running ChatGPT ads, but results have yet to meet expectations. The distinction is between advertisers trying the platform and advertisers committing recurring budgets. Adoption is moving faster than evidence of durable returns. 2/ For Meta, where the money comes from matters as much as how fast ChatGPT grows. The report describes two phases: A. Test budgets: These are generally fixed. Money allocated to ChatGPT can come out of spending that Meta, Google or another platform would otherwise receive. B. Regular budgets: Once ChatGPT becomes an ongoing line item, agencies say clients generally seek additional funding. That would reduce the direct pressure on existing platform budgets. Across the six agencies, Google has the most direct overlap. Traditional US advertisers tend to fund ChatGPT from search budgets, while smaller, more agile companies are more likely to treat it as additional spend. There is an exception worth highlighting: Asia Agency A reports funding coming from several platforms, including Meta. My read: the checks support limited visible impact on Meta so far, but they do not establish that all ChatGPT spending is incremental. The key question is how much test spending becomes recurring spend—and how that recurring spend gets funded. 3/ Performance is mixed, and measurement could change the picture. A. Some encouraging early results: North America Agency A reports good ROI, though volume is limited. The European agency sees ChatGPT working as a complement to search and retail media. B. Weak conversion in other accounts: Asia Agency A reports strong clicks but a sharp drop-off afterward. For one education client, cost per lead was ~1.8x Google’s and well above Meta’s. Asia Agency B reports poor returns and few repeat clients. C. Possible under-attribution: Asia Agency C suspects some users discover products on ChatGPT but convert later through Meta or another channel. This remains an agency hypothesis; better measurement could change how advertisers allocate budgets. FUNDA’s conclusion is that ChatGPT is beginning to affect the ad industry, with spending still mostly incremental or coming from Google. It does not include a ChatGPT drag in its Meta model. My takeaway: ChatGPT is becoming a credible destination for ad budgets, but the competitive impact remains early and uneven. For Meta, the most useful indicators are repeat spending, conversion performance and the source of funding—not advertiser adoption alone. Subscribe to FUNDA Institutional Solution to access more in-depth reports. [email protected].

How does FUNDA’s latest $META Preview analyze ChatGPT’s impact on the advertising industry?

Research|LLM: DeepSeek V4.1-Flash-1017 Reaches the Top Tier of Open-Weight Models, Still Behind Proprietary Frontrunners Faster, lower-cost coding and agent tasks intensify competition DeepSeek released V4.1-Flash-1017 on October 17, an update to the V4.1 Flash model launched on September 10. The API model name is unchanged. Pricing remains $0.30 per million input tokens and $1.20 per million output tokens, with a 50% off-peak discount. The weights are available under the MIT license. The new version outperforms its predecessor on all 11 benchmarks reported by DeepSeek: its Terminal-Bench 4.0 score increased from 31.2 to 42.3 and its DeepSWE score from 74.2 to 77.0. In our view, it now offers the best balance of capability and cost among open-weight models for coding and agent tasks. Since the release, some articles have claimed that 1017 beats GPT-6, mainly on the strength of its DeepSWE score and cost per task. Our preliminary view is that 1017 has joined Kimi K3 and Xiaomi’s MiMo-V2.6 Pro in the top tier of open-weight models. DeepSeek has made clear progress, but this update has not brought the open-weight frontier any closer to the leading proprietary models. Our channel checks indicate that data from the previous generation of proprietary models remains accessible. Distilling from newer models, particularly Fable 5.1 and Opus 5.5, has become more difficult. Opus 5.5 retains a clear lead, and 1017 has not narrowed the gap. OpenAI also has an even stronger internal model, Bel. Detailed Report fundaai.substack.com/p/deepl…

Financing conditions can affect both neoclouds’ cost of capital and their customers’ ability to pay. Neolabs have become important marginal buyers of short-term GPU capacity, paying high rental rates despite uncertain business models. As we flagged on a client call last month, we think their funding could support roughly six to twelve months of spending, after which continued purchases would depend on raising more capital. $CRWV $NBIS $IREN Read more in our weekly post: fundaai.substack.com/p/weekl…

Visibility in semis is several quarters at best. You could argue about $AMD attractiveness as a second source hedge with ASICs in the picture but debating forecasts out to 2030 is not very helpful (in our view).

Thanks @ParadisLabs. We drew some observations from our conversations with people working in the field and published them here: fundaai.substack.com/p/deepp…

In addition to OCP, we also attended the inaugural edition of the iconic Sohn conference in Tokyo. Some Funda readers recognised us and we made new friends. Get in touch with @illyquid for feedback on the following stock pitches: 🔵 Oasis Management - Aoki Holdings 🔵 Indus Capital Partners - Rigaku 🔵 Blue Orca Capital - Rohm / Chubu Electric / TRE Holdings / Toshiba / Kioxia 🔵 Ruane Cunniff - Money Forward 🔵 Pertento Partners - Long Hanwha Corp; Short Ecopro 🔵 SILQ Capital - Makino Milling 🔵 Japan Catalyst - OBIC Business Consultants 🔵 TriVista Capital - Harmonic Drive 🔵 CHANGE Global Investment - Telkom Indonesia 🔵 Polymer Capital - Orix @InvestInJapan @MomokaYokoyama_ @yasutaka_tamura

As we approach AGI, more compute has to go to training, which makes ROI harder to prove, which makes the investing experience less pleasant, as our collective experience this week showed. Read our reaction to the Anthropic ARR number here: fundaai.substack.com/p/colum… Thanks for the shoutout @SaaSletter !

$SMTC sounded confident about the next two years but conservatively guided growth in line with the industry in 2029 and 2030. More importantly, we came away impressed with the team's technical chops.

Column|Anthropic: What the New 3Q Numbers Mean Anthropic disclosed 3Q26 revenue of $16.8Bn during its just-completed pre-IPO investor day. The company had previously disclosed 2Q26 revenue of $11.5Bn and 1Q26 revenue of $4.7Bn. There is speculation that the company guided to $100Bn of ARR by year-end, but we don’t believe it has given explicit guidance. Based on the disclosures and our own tracking over time, we built a full-year ARR and revenue model for Anthropic. Detailed Report fundaai.substack.com/p/colum…

Review| $TSM 26Q3: Revenue Guidance Beats Expectations as Expansion Weighs on Margins TSMC expressed confidence in advanced-node and long-term AI demand during its earnings call. Management said that, alongside discussions with direct customers and their downstream customers, its teams had checked US data center construction and power availability, with encouraging feedback so far. Agentic AI and open-source models are also increasing demand for computing chips. TSMC is responding by adding N3 fabs, departing from its usual practice of stopping capacity additions once a node reaches its target. Detailed Report fundaai.substack.com/p/revie…