5. Market Follower Strategy. It is only after reviewing the competitive landscape

and the competitive strategies of those who are in fact the market leaders (Open AI,

xAi, Google as showcased above) that one can begin to tackle one's own strategic

review as a market follower. As shown below, the firm has to find the right product/ market/channel fit. Using such tools as the 4 Ps and the 5-Forces, it must develop

not one but 3 strategies, its business, competitive, and marketing strategies, by iteration.

DeepSeek could be one such follower, using the same data freely available on the Internet

but serving a much bigger domestic market, the Chinese market, and competing against

its own domestic peers (Alibaba, Baidu, Tencent). See point 8 below profiling DeepSeek.

1 The Strategic Review applied to the AI market

We conduct 3 reviews: a strategic review as shown on this page, and a business and an

operational review as shown on the following page.

Depending on whether the firm is a market disruptor, market challenger, market

incumbent or market follower, the strategic review is conducted as follows:

1. Industry Analysis. The strategic review always begins with an industry analysis. In the

IT industry, almost like clockwork, a new market disruptor emerges every ten years.

Today, that disruptor is of course OpenAI.

2. Market Disruptor Strategy. Here too, the strategic review does not begin with your

firm's business strategy but with that of the market disruptor, namely OpenAI. It  

consists of the following phases:

          . Product. 10 years on after its founding in 2015, OpenAI has successfully come up

            with a marketable product, comprised of  a program (the large-language model or LLM),

            a database covering everything publishable online (GPT), and a chatbot (ChatGPT).

            This first phase can be called the core product development or just product phase

          . Go-To-Market Financing. It successfully clinched already in 2019 $1B from Microsoft

            to finance its product development. Killing two birds with one stone, by 2023 it received

            another $12B from them to finance its commercial development. The tie-up allows it

            at the same time to tap into Microsoft's customer base of 1.6B users and 2M businesses.

            This second phase can be called the go-to-market product financing or money phase

          . Standards. As in any industry, the 3 to 5 major players (OpenAI, Anthropic, Meta,

            Google, DeepSeek) must agree on an industry-wide standard, which is the only way

            for individual firms to benefit from economies of scale and bring down costs. This is

            why OpenAI and Anthropic are pushing to have their LLM models, respectively

            o3 and MCP, adopted as the standard for AI, just as Linux became the standard

            for operating systems, to be distributed for free to lure the greatest number of users

            This third phase can be called the scaled product development or industry phase

          . Partnerships. OpenAI has recently acquired for $6.7B Jony Ive's design firm iO, in

            order to fill a missing gap in hardware, to lure this time the greatest number of buyers

            This fourth phase can be called the differentiated product development or market phase

3. Market Challenger Strategy. A challenger to watch closely is Elon Musk's xAI. Just

as Apple was also a challenger until it bypassed in a few short years Blackberry, Elon has

very quickly narrowed OpenAI's lead by  releasing his own AI engine in Grok, his own

AI chip to equip his Dojo supercomputer, and a critical mass of 1.6B users he can tap

into at X and now at Telegram, both to exclusively use Grok. Like OpenAI, he has

delivered to his investors, who put in $6B, on the product, the market, and the channel.

INTERNET NETWORK

LANGUAGES

CLOUD OF THINGS

DATA

OPERATING

SYSTEMS

CLOUD SOFTWARE

SECURITY SOFTWARE

NETWORK SOFTWARE

APPLICATION SOFTWARE

HARDWARE

& RELATED

SOFTWARE

DATA ANALYST

APPLICATION DEVELOPER

ANALYTICS & BI

EXPERIENCE

DESIGNER

SHARED WORKSPACE

WORKFLOW

ERP

DIGITAL CURRENCIES

MARTECH

Competitive Strategies

OPEN SOURCE FRAMEWORK

EXPERIENCE DESIGN

STANDARDS

HARDWARE & SOFTWARE

DEVELOPER TOOLS

DIGITAL EDUCATION

Social Media & eCommerce

TELCOS (MSOs)

DATA CENTER OPERATORS

WEB PLATFORMS

Market Disruptor

How We Do It

Strategy

4. Market Incumbent Strategy. OpenAI backed by Microsoft, Anthropic backed by

Amazon,  and xAI backed by Elon must face off the most formidable incumbent that

is Google. As shown on the 2-hour-long video of its 2025 Keynote below, the search

giant has made great strides, having nearly finished making end-user tools which the 3

upstarts do not yet have. To go against OpenAI's AI engine ChatGPT, Anthropic's Claude,

and xAI's Grok, Google must have easily matched the 3 upstarts' $10B to $15B of capital commitments, in order to develop and roll out its own AI engine, Google Gemini.

To understand the extent of Google's formidable reach, let us not forget that the founders

of OpenAI, Anthropic and now Safe Superintelligence, Inc (SSI) cut their teeth in their

twenties as crack developers at Google's in-house AI incubator, Google Brain (notably Ilya

Sutskever, later chief scientist at OpenAI and after a falling out with Sam Altman now at

his own firm, Safe Superintelligence, and the Amodei siblings at Anthropic).

However one wants to cut it, the Internet business we've now rebranded as AI is still

Google's "sort and search" business extended now as a "sort, search, and solve" business

(Google also holds an equity stake, albeit a small one, in Anthropic, having pledged to commit a total

of $2B to Amazon's $8B as of end 2024)

AI Disruptor

AI Disruptor

Competitive Advantage

The Strategies

Segmentation, Positioning, Differentiation

(4 Ps, 5-Forces)

see Meta Case Study

Product, Market, Channel

see NFTs Case Study

Market Challenger

Market Incumbent

Market Follower

Industry Analysis

See also The Internet page

7. China. Finally there is the biggest threat of all, China. As we saw above with the

top US market leaders, the 4 critical success factors are in the product's software

(AI engine), and hardware (AI chip), the market's billion+ users, and the financial

backing. China has all four. On the software side, the top Chinese Internet companies,

such as Alibaba, Tencent, Baidu, and an OpenAI clone, DeepSeek, are already at

the 3rd version of their AI engines, respectively Qwen 3, HPI, Ernie 4.5 and

DeepSeek. On the hardware side, Huawei and Xiaomi, to name just two we know

about, have already launched their AI chips, respectively the 910C GPU and the

XRing01. In terms of market  size, Tencent's Wechat alone has a base of 1.3B users.

The Money. Then there is the invisible hand of the Chinese government. As it has

done with EVs, it has already shelled out this year alone some $60B of subsidies to

the Chinese IT industry (our governments have also subsidized as much but without

any strings attached, throwing good money after bad).

The Brainpower. That the Trump Administration has pushed forward its $500B

Stargate project with Oracle, OpenAI and Softbank and forced Apple to manufacture

at home is a good sign. The question now is how the US will train 400 000 new

engineers a year to catch up to the 600 000 the Chinese churn out of their now

world-class engineering schools.

8. DeepSeek. Below is a profile of DeepSeek's founder, Liang Wenfeng, a multi-

talented entrepreneur who, after completing his master's in computer science in 2010,

developed already back then an AI stock-trading program to take advantage of the

capital markets' instability and volatility after the 2008 GFC. Capitalizing on its

success, he created his own hedge fund, raising up to 19B yuan. In a sense, Liang

was already stress testing what was to become DeepSeek. He applied it to the most

difficult usecase there was, using the same type of AI algorithmic trading we had

at home, to invest in the Chinese stock market for his clients.

It therefore came as no surprise that Liang  took a step further in 2019 by founding

High Flyer AI to develop what was to become DeepSeek. When the first R1 version

of DeepSeek was rolled out in January 2025, upon learning that its development

had cost a mere $6M compared to the tens of billions of dollars raised by OpenAI,

US tech stocks tumbled, losing $1T in market value.

Moreover, as the graph above "The Rise and Rise of AI" shows, the Chinese IT

industry didn't wait for DeepSeek's or OpenAI's roll-outs in 2024 and 2025 to begin

investing heavily to realign itself to AI. As shown on the graph, its first GPT database,

Wu-Dao 2.0, was released in 2021, preceding  Google's own GLaM and OpenAI's GPT 4.

See the Wikipedia article on Wu-Dao 2.0, an entirely state-sponsored project.

6. Business Models. Then only, after reviewing their respective competitive strategies,

do they develop their business models. The map below, shows who the biggest players

are, as reflected by the size of the circles (Open AI in purple, Google in yellow and

the Chinese in orange), operating their AI data centers powered by their AI chips and

AI databases and running their AI engines. So far, as their own chips are still under

development, only Nvidia GPU chips are used.

Perhaps more importantly, the map also tells us what their distinct business models

would be. To begin with, using the 4 Ps, in terms of the Product, they are tailoring

and pricing their software, the AI engines, to meet the needs of their target markets.

For example, Google has on the one hand their top line Vertex AI LLM, priced higher

to process resource-intensive apps for those of their Fortune 500 customers needing

not just text but images and videos and on the other hand, their Coden, Imagen, and

Chirp LLMs, which are customized into more narrowly defined apps and priced lower.

In terms of the remaining 3 Ps, Price, Place and Promotion, nothing has changed.

They will continue to use dual pricing, billing a fixed subscription fee and a variable

usage fee, based on the number and complexity of the queries known as tokens. As

for the marketing channel downstream, because they target the entire total addressable

market or TAM, they will partner with the usual intermediaries, the wholesalers and

the retailers. We will therefore continue to have a Kyndril, an IBM spinoff, to operate

the data centers and IBM itself to provide the higher margin installation services.

Finally, regarding their promotional strategy, new AI Martech tools have yet to be

developed with regard to online ad auctions, PR events, Github-like developer websites,

AI app stores, and the like.

Table of Contents

The Strategy Process

applied to the AI market

The Product:

1. Industry Analysis

The Product/Market/Channel Fit:

2. Market Disruptor Strategy

3. Market Challenger Strategy

4. Market Incumbent Strategy

5. Market Follower Strategy

6. Business Models

7. China

8. DeepSeek