Futures
Access hundreds of perpetual contracts
CFD
Gold
One platform for global traditional assets
Options
Hot
Trade European-style vanilla options
Unified Account
Maximize your capital efficiency
Demo Trading
Introduction to Futures Trading
Learn the basics of futures trading
Futures Events
Join events to earn rewards
Demo Trading
Use virtual funds to practice risk-free trading
CFD
Stock CFD Derivatives
US Stocks
Access real US stocks and ETFs
HK Stocks
Trade quality Hong Kong-listed stocks
Korean Stocks
SK Hynix
Real Korean stocks and top assets
Stock Futures
High leverage, 24/7 trading
Tokenized Stocks
Backed by real stock assets
IPO Access
Unlock full access to global stock IPOs
GUSD
3.8%
Mint GUSD for Treasury RWA yields
Stocks Activities
Trade Popular Stocks and Unlock Generous Airdrops
Launch
CandyDrop
Collect candies to earn airdrops
Launchpool
Quick staking, earn potential new tokens
HODLer Airdrop
Hold GT and get massive airdrops for free
Pre-IPOs
Unlock full access to global stock IPOs
Alpha Points
Trade on-chain assets and earn airdrops
Futures Points
Earn futures points and claim airdrop rewards
Promotions
AI
Gate AI
Your all-in-one conversational AI partner
Gate AI Bot
Use Gate AI directly in your social App
GateClaw
Gate Blue Lobster, ready to go
Gate for AI Agent
AI infrastructure, Gate MCP, Skills, and CLI
Gate Skills Hub
10K+ Skills
From office tasks to trading, the all-in-one skill hub makes AI even more useful.
After the semiconductor adjustment, the 8 major AI application sectors that are truly worth paying attention to are
Recently, semiconductors have been falling quite badly, and many people have started to worry whether the AI boom is over. I said earlier that this round of AI will inevitably see a bubble—during the fast development of any technological revolution, the valuation expansion and bubble phase are unavoidable. But the bursting of a bubble doesn’t mean the industry is over; the internet is the best example.
So today, we’ll use history as a mirror—go back to before the internet bubble in 2000, look at the industry’s development path back then, and compare it with which stage today’s AI has reached.
The 2000 internet wave went through about five years of development (1995-2000). The whole industry can basically be divided into six stages:
First stage: Infrastructure build-out (1994-1996)
When the internet was just starting out, the biggest bottleneck back then was network speed. So the first breakout was communication equipment and network infrastructure construction. Companies like Cisco and Lucent became the biggest winners at the time.
Second stage: Upstream hardware (1996-1998)
Once the network was built, the next step was large-scale development of servers, PCs, and chips. Intel, Dell, Sun, and others began to grow rapidly.
Third stage: Software platforms (1997-1999)
Enterprises began fully digitizing. Various databases and enterprise software saw explosive growth. Software companies like Oracle, Microsoft, and SAP saw their valuations rise quickly.
Fourth stage: Internet platforms (1998-2000)
Real internet companies started to rise. Portals and e-commerce platforms—Yahoo, Amazon, eBay, and more—along with search engines, emerged one after another. The market began shifting from “selling shovels” to “people who truly make money by using the internet.”
Fifth stage: Mass internet (1999-2000)
The internet entered a phase of mass frenzy. As long as a company’s name ended with “.com,” the market was willing to give it high valuations. A large number of companies that had no profits and even no products went public.
Sixth stage: Bursting of the internet bubble (2000-2002)
Valuations returned to reality. Many internet companies went bankrupt, but the truly excellent ones survived and later grew into the giants over the following dozen-plus years.
So, which internet stage is AI in today?
I believe that AI has already moved into the transition from the third stage to the fourth stage.
First stage: AI infrastructure—basically completed.
The biggest winners this time are the “shovel sellers” of the AI era.
Including:
* Nvidia (GPUs)
* TSMC (advanced packaging)
* ASML (EUV lithography machines)
* SK hynix (HBM)
These companies provide the most core AI infrastructure.
At the same time, global AI data centers are still being built continuously, but from the perspective of the capital markets, this part has already gone through the first round of valuation increases—so I think the first stage is basically done.
Second stage: AI upstream hardware—also basically done.
With the surge in computing power demand, all kinds of hardware companies around AI data centers have almost gone crazy making money.
For example:
* Power: Fluence (FLNC), Bloom Energy (BE)
* High-speed interconnect: Credo (CRDO), Astera Labs (ALAB)
* Optical communications: Marvell (MRVL)
* Sub-segments like liquid cooling, CPO, high-speed optical modules, and more
I’ve analyzed these directions for everyone before. In essence, they’re all part of the AI data center industry chain. So I think the second stage has also completed the major valuation uplift.
Third stage: AI platforms—becoming the new main character.
Next, what the market begins to focus on is no longer GPUs, but platform companies that have AI capabilities.
Including:
* OpenAI
* Anthropic
* Gemini (Google)
* xAI
* Meta Llama
You can observe a very interesting phenomenon in the recent market trend:
Semiconductors start to adjust, while platform-type companies like Meta and Google keep setting new highs.
Capital has begun to flow from the infrastructure layer into the platform layer.
Fourth stage: AI applications—just getting started.
If the internet back then corresponded to Amazon, Google, Facebook—truly profitable platforms—then the biggest opportunity for AI in the future is very likely in the application layer.
AI applications are still at an early stage, but some directions worth focusing on are already starting to emerge.
---
Fifth stage: Mass AI.
In the future, when every company starts emphasizing that it is an AI company, and every product tags “AI,” it’s time to be wary of a truly big bubble.
Back then, the internet was the “.com” era. In the future, it might very well be the “AI+” era.
So, where do I think AI is positioned right now?
My view is:
Today, AI is roughly in the 3rd-4th stage of the internet development path—when platform layer growth is taking off and the application layer begins to take over.
If this judgment holds, then in the next few years, the focus likely won’t be simply on chasing GPUs. It will be about finding application companies that can truly make money by using AI.
So, what AI application-layer companies are worth watching right now?
I roughly divide the currently listed U.S. stocks into several categories.
First category: AI office
1、Microsoft (Microsoft)
Copilot has already been fully integrated into Office, Windows, GitHub, and Azure. In the future, Microsoft could very likely become the biggest beneficiary of enterprise AI office.
2、Salesforce (CRM)
The world’s largest CRM software company is currently fully advancing Agentforce and rolling out AI sales, AI customer service, and AI marketing.
3、ServiceNow (NOW)
A leader in enterprise automation software. After introducing AI, it is putting a large amount of approvals, IT operations, and enterprise workflow automation into place.
Second category: AI programming
1、GitHub (Microsoft subsidiary)
Copilot has already proven that developers are willing to keep paying for AI programming tools.
2、Cursor
Not yet publicly listed, but it is one of the fastest-growing companies in the AI coding space, and is worth focusing on.
3、Cognition (Devin)
A representative company for AI programmers. It is also not publicly listed.
Third category: Enterprise AI (B2B)
1、Palantir (PLTR)
One of the hottest AI application companies this round. Many people call it an “enterprise AI operating system.”
2、Snowflake (SNOW)
An enterprise data platform. In the future, it will be an important entry point for AI to read enterprise data.
3、Datadog (DDOG)
In the AI era, lots of log analytics, performance monitoring, and system observability are needed. That’s why Datadog is also worth watching.
Fourth category: AI design
1、Canva (not publicly listed)
2、Adobe (ADBE)
Generative AI like Firefly is continuously improving design efficiency.
---
Fifth category: AI search
Google (GOOGL)
Gemini has been fully integrated into Google Search’s ecosystem, and is an important representative of AI search today.
Sixth category: AI healthcare
1、Recursion (RXRX)
An AI drug discovery company invested by Nvidia, focused on drug development.
2、Schrödinger (SDGR)
Uses AI for molecular simulation and drug discovery.
Seventh category: AI finance
Robinhood (HOOD): is gradually introducing AI capabilities into financial services like investing and wealth/financial advisory.
Eighth category: AI Agent (intelligent agents)
This is the direction I think is most worth watching in the next few years.
Many of the current leaders are not publicly listed yet, including:
* OpenAI
* Anthropic (Claude)
* Perplexity
* Harvey
* Glean
* Cursor
If these companies eventually go public one after another in the future, they could very likely become the most important new core for the next round of the AI cycle.
To summarize my view:
The biggest winner in the internet wasn’t the company that sold network equipment first. It was the platform and application companies that later truly used the internet to create business value.
AI may follow the same path.
Infrastructure will still matter, but the market’s focus is likely to gradually shift from hardware like GPUs, HBM, and optical modules toward AI platforms and AI applications that can truly create sustained cash flow.
So what I think is most worth researching next isn’t who can still sell more GPUs, but who can truly use AI to build a business model and achieve profitability. That may be the biggest investment opportunity in the second half of the AI era.
$PLTRX