$1($b$)$ In addition to its data centre in Texas, BitTech currently has $1,000$ servers located in Singapore. These are deployed for cloud services. With the advent of AI technologies such as Google Bard, OpenAI chatGPT and Microsoft coPilot, BitTech is thinking about pivoting from cloud services to AI services. It is evaluating whether it should pivot $500$ servers to the new AI services. These $500$ servers have a remaining shelf$-$life of $7$ years. At the end of this period, all these servers will be written$-$off with zero book value. Presently in Singapore, monthly sales per server are $$24,000.$ Operating costs per server are $$5,000$ a month. It expects sales of cloud services to decline by $5\%$ each year. Costs per server $($for cloud services$)$ are expected to increase by a gradual $3\%$ per year. On the other hand, monthly sales per server for AI services are estimated to be $$36,000.$ However, operating costs per server are higher at $$10,000$ per month. Sales of AI services are expected to increase by a hefty $15\%$ each year. The costs of operating such servers will also increase by $12\%$ per year.

For all its projects, BitTech uses a minimum ROI target of $18\%.$

$($i$)$ Create an appropriate NPV$-$based model to analyse whether BitTech should pivot to AI services for $500$ of its existing cloud servers. State any two $(2)$ assumption$($s$)$ used by your model. $(12$ marks$)$

$($ii$)$ Using your model, examine whether BitTech should pivot to AI services. Explain why. $(4$ marks$)$

Note to expert: Please answer $1($b$)$ provide the models in excel table form. Not in written words.Question $1$

BitTech Pte Ltd $($BitTech$),$ with headquarters in Singapore, operates a computer server farm in

Texas, USA. It has some $4,000$ servers mining Bitcoin cryptocurrency. However, with Bitcoin

values fluctuating wildly, intense competition and increasing operating costs, the company has

decided to pivot its business to cloud services and to relocate its servers to Asia.

Table $1$ below shows technical data over the last $6$ months in $4$ countries that BitTech is

evaluating for the suitability of migrating its servers.

Table $1.$ Technical Data for Server Operation in $4$ Countries

The table above shows the electrical voltages that the power supply companies in the various

countries can provide. Availability indicates how many hours in a month where there is no

power outage. BitTech works on the basis of $365$ days a year over $12$ months. Some fluctuations

in voltage values and availability days are normal. The table shows measured values over a $6-$

month period.

The servers are sensitive electrical equipment and require stable power from the data centre

that houses the servers. To operate reliably, the data centres used by BitTech must be rated at

$22kV$ with a tolerance of $6\%.$

As an example, consider a data centre that is rated at $10kV$ with a tolerance of $10\%.$ This

means that it can only operate the servers if the power supplied varies between $9kV$ and $11kV$

with a long$-$term average of $10kV.$

BitTech also noted that the electricity rates in Indonesia, Malaysia, Singapore and Vietnam are

$$0\mathrm{.}16,$$$0\mathrm{.}14,$$$0\mathrm{.}25$ and $$0\mathrm{.}15$ per $kWh,$ respectively. These rates are expected to remain

relatively unchanged for the foreseeable future.

BitTech prides itself of its ability to provide services at a guaranteed availability of at least

$99\mathrm{.}95\%.$ For added assurances, it also requires the power suppliers to match this level of

availability.

In the industry, the Percentage Availability metric is defined as System Available Hours $/$ Total

Calendar Hours expressed as a percentage. Thus, if a system is available for service for $657$

hours in a month, the Percentage Availability would be computed as $90\%$ per month.

$[$Hint: $90\%=0\mathrm{.}90=657$ hours $/(24$ hours $x365$ days $/12$ months$)]$