Upwind, a San Francisco cloud security company, announced a $400 million Series C on October 28, 2026, at a $3.8 billion valuation. TechCrunch reported that Bessemer Venture Partners led the round, with existing backers Greylock Partners and Craft taking part. The Upwind Series C lifts total funding past $800 million, according to the company’s own announcement.
The speed matters as much as the size. Upwind says the new money arrived less than nine months after its $250 million Series B, which TechCrunch pegged at a $1.5 billion valuation. That works out to a valuation multiple of roughly 2.5x in under a year, a pace few infrastructure software companies manage in a tight funding market.
Cloud security is absorbing capital because cloud spending itself keeps compounding. Microsoft’s cloud arm just posted 48% Azure growth to $33.2 billion, and every dollar of that workload growth creates new attack surface. This analysis covers who funded Upwind, what the platform does, what is still unconfirmed, how the competitive field looks, and where the funding points next.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
What Upwind Announced on October 28
The headline facts are simple. Upwind, which also trades under the name Upwind Security, closed a $400 million Series C at a $3.8 billion valuation. Bessemer Venture Partners led. The company describes its product as a cloud and AI security platform built to protect cloud and AI infrastructure.
Chief executive Amiram Shachar framed the raise in a post on the company’s site. He wrote: “This additional capital will fuel our product and engineering growth in building the platform that will secure cloud & AI infrastructure for the world in the AI era.” That sentence is the clearest statement of intent in the announcement. It points the money at engineering and product, not at a sales blitz or an acquisition spree.
Coverage landed fast. SecurityWeek ran the valuation step-up from the Series B, and Bloomberg Law carried the wire story under an AI cybersecurity framing. Upwind itself keeps the primary record in its news archive. All of them agree on the three numbers that matter: $400 million, $3.8 billion, October 28.
Who Funded the Upwind Series C
Bessemer Venture Partners leads, which gives the round a recognizable anchor. The announcement also names a long participant list, and the length of that list says something. Late-stage security rounds often come with two or three funds. This one drew a dozen.
| Investor | Role in the Series C |
|---|---|
| Bessemer Venture Partners | Lead investor |
| Greylock Partners | Named participant |
| Cyberstarts | Named participant |
| Craft | Named participant |
| Salesforce Ventures | Named participant |
| Alta Park | Named participant |
| Picture Capital | Named participant |
| Swish Ventures | Named participant |
| Cerca Partners | Named participant |
| TCV | Named participant |
| Penny Jar | Named participant |
| Leaders Fund | Named participant |
| Canapi ventures | Named participant |
Source: Upwind’s announcement. The company did not publish a per-investor breakdown of check sizes, so none is shown here.
Why the cap table matters
Salesforce Ventures stands out because it is a strategic investor with its own enterprise customer base. A strategic name on a security cap table can open distribution doors. Upwind has not said what, if anything, that relationship covers, so treat it as a signal and not a commercial agreement.
Repeat backers tell a second story. Greylock, Cyberstarts, Craft, Alta Park, TCV and Leaders Fund all appear in the announcement, and Upwind’s own archive ties several of them to earlier rounds. When insiders keep writing larger checks at higher prices, they are betting on the trend line, not on a pitch deck.
From $1.5 Billion to $3.8 Billion: The Valuation Jump
The valuation arc is the sharpest part of the Upwind funding story. According to TechCrunch, the Series B valued the company at $1.5 billion. The Series C sets it at $3.8 billion. By simple arithmetic, that is an increase of $2.3 billion, or about 153%, in less than nine months.
| Round | Amount | Valuation | Timing |
|---|---|---|---|
| Series A | $100 million | Not stated in sources used here | December 2024, per Upwind’s news archive |
| Series B | $250 million | $1.5 billion (TechCrunch) | Less than nine months before the Series C |
| Series C | $400 million | $3.8 billion | October 28, 2026 |
| Total raised | More than $800 million | n/a | Per Upwind’s announcement |
The round sizes also climb in a clean staircase: $100 million, then $250 million, then $400 million. Add those three and you reach $750 million, which means roughly $50 million or more sits in earlier financing that the sources do not itemize. That is simple addition, not a reported figure, and it only reconciles the “more than $800 million” claim.
One more calculation helps frame dilution. If $3.8 billion is a post-money number, a $400 million raise equals about 10.5% of the company. The sources do not say whether the valuation is pre- or post-money, so that percentage is an illustration and not a disclosed stake.
What the Upwind Cloud and AI Security Platform Does
Upwind sells a platform that secures cloud and AI infrastructure. In the security industry, that category usually gets called a cloud-native application protection platform, or CNAPP. These products scan cloud accounts, containers, Kubernetes clusters and workloads, then rank what to fix first.
Search coverage of the announcement describes Upwind’s angle as runtime-native security. The idea: watch what workloads actually do while running, instead of only scanning configurations at rest. Runtime context helps teams cut alert noise, because a flaw in code that never executes matters less than one in a service handling live traffic. Reports also say the platform pulls signals from workloads, networks, APIs, data flows and AI agents. Upwind’s own materials are the place to check exact feature scope, and buyers should request a proof of concept instead of trusting a summary.
Where runtime data beats static scanning
Static scanners are good at finding misconfigurations early. They struggle to tell which of 10,000 findings an attacker can actually reach. Teams that already run open-source scanners will recognize the problem from our guide to five vulnerability scanner tools, where raw findings pile up faster than staff can triage them. Runtime tools try to shrink that pile by adding “is it live and exposed” to every ticket.
The Datadog Question: What Is and Is Not Confirmed
TechCrunch’s headline frames the round around Upwind walking away from a Datadog acquisition offer. That framing has circulated widely. It is also the one part of the story the official announcement does not back up. Upwind’s post covers the Series C, the investors and the roadmap. It says nothing about an offer from Datadog or any other buyer.
For this article, the Datadog angle is unconfirmed. Neither company has published terms, a price, or a timeline in the sources reviewed, and we will not repeat details we cannot verify. If either side confirms the approach on the record, that would change the read on the valuation, because a rejected bid sets a floor under what a buyer thought the company was worth.
The safest reading is narrow. Upwind chose to raise $400 million as an independent company at $3.8 billion. What it declined, and from whom, remains a reported claim.
Why Cloud Security Funding Keeps Climbing
Investors are not chasing a fad here. Security budgets are growing, and cloud security software takes a large share of the increase. Gartner’s July 2025 forecast put worldwide end-user spending on information security at $213 billion in 2025, and the same release projects 12.5% growth to $240 billion in 2026.
| Gartner metric | Figure | Year |
|---|---|---|
| Worldwide end-user information security spending | $193 billion | 2024 |
| Worldwide end-user information security spending | $213 billion | 2025 (projected) |
| Worldwide information security spending | $240 billion (+12.5%) | 2026 (projected) |
| Worldwide security software spending | $94.96 billion | 2024 |
| Worldwide security software spending | $105.94 billion | 2025 (projected) |
The release states: “Worldwide end-user spending on information security is projected to reach $213 billion in 2025, up from $193 billion in 2024.” A second line from the same Gartner material reads: “Worldwide information-security spending is projected to increase 12.5% in 2026, reaching $240 billion.” And on software specifically: “Worldwide security-software spending is projected to rise from $94.96 billion in 2024 to $105.94 billion in 2025.”
Those numbers set the backdrop. Software spend grew about 11.6% from 2024 to 2025 by our math ($94.96 billion to $105.94 billion), so the pool Upwind sells into is expanding at a double-digit clip. A company that quadruples revenue, as Upwind’s funding press release headline claims for 2026, would be taking share from a market growing well under half that speed. Upwind has not published the revenue base, so the multiple says little about absolute size.
Competitive Picture: Where Upwind Sits in the CNAPP Race
Upwind does not operate in open space. Reporting on the round names Wiz, Palo Alto Networks, CrowdStrike and Orca Security as relevant competitors in cloud security. The search coverage reviewed for this article gives no current market-share or revenue figures for any of them, so a numbers-based ranking would be guesswork. Here is a qualitative read instead.
- Platform incumbents such as Palo Alto Networks and CrowdStrike sell cloud security next to firewalls, endpoint tools and identity products. Their pitch is one vendor and one contract. Their weakness is that cloud-first startups often ship faster inside a single domain.
- Cloud-native specialists such as Wiz and Orca Security built around agentless scanning and a security graph. Upwind competes with them by leaning on runtime signals.
- Observability and monitoring vendors like Datadog already sit inside production workloads. That proximity explains why the unconfirmed acquisition angle sounds plausible to many readers, even without a confirmation.
The competitive question for buyers is less “which logo wins” and more “which approach cuts the alert queue.” A startup with $400 million can hire hard-to-find cloud engineers and buy distribution. It still has to prove, customer by customer, that its detections beat the incumbents already wired into the enterprise contract.
AI Infrastructure Changes the Security Math
Upwind chose to put AI in its category name, and that is a deliberate bet. Companies are racing to deploy models, agents and inference pipelines on cloud infrastructure, and those workloads bring new risks: exposed API keys, over-permissioned agents, and data flows that never appeared in old architecture diagrams. Shachar’s quote about securing “cloud & AI infrastructure for the world in the AI era” speaks directly to that shift.
The capital side supports the thesis. Our report on the AI cash crunch and Meta’s $95 billion search shows how much money is flowing into compute. Every new data center and neocloud needs a security layer, and buyers of that layer have little time to build it themselves.
The attacker side is moving as well. When we covered GPT-6.1 Terra’s critical cyber tier, the point was that capable models now cost about $1 per million tokens. Cheap model access lowers the price of automated reconnaissance. That is a tailwind for any vendor selling detection that runs continuously instead of on a weekly scan schedule.
Historical Context: 22 Months From Series A to Series C
Upwind’s own news archive dates its $100 million Series A to December 2024, led by Craft with TCV and Alta Park joining existing investors. From that round to the Series C on October 28, 2026 is 22 months. In that span the company went from a $100 million raise to a $400 million raise, and from a valuation not disclosed in our sources to $3.8 billion.
Fast compounding like that rarely happens in a vacuum. It usually follows a few things at once: a product that wins head-to-head trials, a market where the buyer has budget, and investors who fear missing the category leader. The Series B to Series C gap of under nine months suggests all three were in play.
The cloud itself has gone through the same cycle. Enterprises first moved workloads, then discovered misconfigurations, then bought posture tools, then wanted runtime protection. Breaches keep reinforcing the lesson. The TiDB Cloud breach through a stolen service account lasted 28 days, which is exactly the type of dwell time that runtime monitoring aims to shorten.
What $400 Million Buys: Product, Engineering and Reach
Shachar’s wording prioritizes product and engineering growth. Reports on the round add that Upwind intends to widen coverage toward endpoints, code and cloud environments, and to expand into identity, data and AI security. Those details come from coverage of the announcement and not from a published roadmap, so timelines remain open.
Widening scope is the standard move for a security company at this stage. Once a vendor owns the cloud workload view, the next logical step is stitching in identity and data context, so one alert explains who did what to which asset. It is also the move that puts Upwind directly against the platform incumbents, which already sell those adjacent products.
Hiring is the real bottleneck
Money alone does not ship detections. Security engineering talent for Kubernetes, eBPF-style runtime sensors and machine learning is scarce, and every funded rival is bidding for the same people. A $400 million balance sheet gives Upwind room to outbid on equity and salary, which may matter more than any marketing budget over the next 12 months.
What Security Teams Should Do Now
A funding round is not a product review. Still, large rounds change vendor behavior, so buyers should adjust their plans accordingly. Here is a short checklist:
- Ask for a runtime proof of concept. Run it against your noisiest cluster and count how many alerts disappear versus your current scanner.
- Check your software supply chain data. If you generate bills of materials already, compare them against what any platform reports. Our rundown of Syft, Trivy and cdxgen shows what a baseline looks like.
- Demand pricing clarity. Well-funded vendors sometimes discount hard to win logos, then reprice at renewal. Get multi-year terms in writing.
- Map AI workloads. List every model endpoint, agent and vector store in your cloud accounts before you evaluate any tool that claims AI coverage.
Teams running managed Kubernetes should also revisit cost and architecture choices that affect the security surface. Our comparison of EKS versus ECS covers one of those trade-offs. For wider reading, the cloud computing section collects related guides.
Risks and Open Questions
A $3.8 billion valuation sets a high bar. To justify it, Upwind must keep growing at a pace that outruns a security software market projected by Gartner to grow roughly 12% a year. The company says revenue quadrupled in 2026, which is an eye-catching figure, but the sources do not disclose the revenue base, customer count or net retention.
Several unknowns remain:
- Whether the Datadog approach reported by TechCrunch happened, and on what terms.
- Whether the $3.8 billion figure is pre-money or post-money.
- How much of the $400 million is primary capital versus any secondary sale to existing holders. The announcement as summarized in the sources does not say.
- How the platform performs against incumbents in independent tests.
- When the promised identity, data and AI security modules will ship.
None of these undercuts the round. They simply mark the places where careful readers should wait for evidence before drawing hard conclusions.
Five Predictions for Cloud and AI Security
These are our own forecasts, not statements from Upwind or its investors. Each rests on the facts above.
- More $300 million-plus security rounds will follow. When a leader in a hot category closes this big, rivals and adjacent startups raise to keep pace. Expect at least a few more late-stage cloud security rounds before mid-2027.
- Runtime context becomes a checkbox feature. Incumbents will market runtime detection as standard. Upwind will have to win on depth, not on the category label.
- AI agent security gets its own budget line. By the next planning cycle, large buyers will likely separate spending for agent permissions and model endpoints from general cloud security.
- An IPO or large acquisition conversation arrives within two years. With over $800 million raised and a $3.8 billion mark, the exit options narrow to a public listing or a strategic buyer. The reported Datadog interest, if true, shows buyers are already looking.
- Pricing pressure builds on posture-only tools. As runtime platforms prove they cut alert volume, buyers will resist paying premium rates for scanners that only list problems.
FAQ: Upwind’s $400 Million Series C
How much did Upwind raise and at what valuation?
Upwind raised $400 million in a Series C at a $3.8 billion valuation, announced on October 28, 2026.
Who led the Upwind Series C?
Bessemer Venture Partners led the round. Greylock Partners, Cyberstarts, Craft, Salesforce Ventures, Alta Park, Picture Capital, Swish Ventures, Cerca Partners, TCV, Penny Jar, Leaders Fund and Canapi ventures also took part, per Upwind’s announcement.
How much has Upwind raised in total?
More than $800 million, according to the company’s announcement. That includes the $250 million Series B from earlier in 2026.
What does Upwind do?
Upwind builds a cloud and AI security platform for protecting cloud and AI infrastructure. Coverage of the round describes a runtime-focused approach that uses live workload context to prioritize risks.
Did Upwind reject a Datadog acquisition offer?
TechCrunch’s headline says so, but Upwind’s official announcement does not mention an offer. Treat the claim as unconfirmed until either company comments on the record.
What will Upwind do with the money?
CEO Amiram Shachar said the capital will fuel product and engineering growth. Reports add plans to broaden coverage across endpoints, code and cloud, and to expand into identity, data and AI security.
Who competes with Upwind?
Coverage names Wiz, Palo Alto Networks, CrowdStrike and Orca Security as relevant competitors in cloud security. No current market-share data for them appeared in the sources reviewed.
Is this a good sign for the cloud security market?
It shows strong investor appetite. Gartner projects worldwide information security spending to reach $240 billion in 2026, up 12.5%, and a round of this size suggests investors expect cloud security to take a growing slice of it.
